LED Bulbs

[Jerry_NJ]

I have a problem with the statement: "white LED-based lights are still less efficient than current CFLs", given what I know about LEDs. Talking about equivalent light to tungsten bulbs (the standard) a 40W CFL runs in the 10 watt consumption range, while the LED is 1.5 watts. Now I agree the CFL 40W puts out more light to my eye than does the 40W LED, but not 6 times more light.

Also, a more complex power supply: step-down voltage and full wave bridge rectifier would increase the efficiency of the LED light on an A/C circuit. But, this complex power supply will have some of its own losses, will be bulky (by comparison) and will add serious cost.

Given the low cost of a full-wave bridge and a resistor, an LED setup that uses enough LED in series to approach 100 volts to light-up full should be very efficient, albeit a fixed brightness. This would be about 25 LEDs which is still not very bright in my experience.

I'm not sure if more LEDs adds more spread of light, or more intensity, or both. This suggests for brighter LED fixtures two banks of 25 LEDs, in this example, which has two banks in parallel each has its own rectifier and resistor thereby eliminating the parallel "problem". The two banks are in effect each connected to the A/C source.

 

[DiscoInferno]

I have a problem with the statement: "white LED-based lights are still less efficient than current CFLs", given what I know about LEDs. Talking about equivalent light to tungsten bulbs (the standard) a 40W CFL runs in the 10 watt consumption range, while the LED is 1.5 watts. Now I agree the CFL 40W puts out more light to my eye than does the 40W LED, but not 6 times more light.

From the DOE:
http://www.netl.doe.gov/ssl/PDFs/LED-FAQ.pdf
"The best white LEDs are similar in
efficiency to CFLs, but most of the
white LEDs currently available in
consumer products are only marginally
more efficient than incandescent lamps.
Lumens per watt (lpw) is the measure
of how efficiently the light source is
converting electricity into usable light.
The best white LEDs available today
can produce about 45-50 lpw. For
comparison, incandescent lamps typically
produce 12-15 lpw; CFLs produce at
least 50 lpw."

I agree this will not be the case for long, but CFLs have a significant head-start in terms of commercialization.

 

[Jerry_NJ]

Hum, thanks, I am surprised at the nubmers for LEDs vs CFL.

Perception, perception.... is reality in many ways. Regardless of what the DOE says, never mind they are an (incompetent?) government agency, I see and have applications for LED lights in 120vac fixtures that save a high percentage of electrical energy as measured against CFL. One is as I stated in an earlier post, and application in the basement that is dark day and night and has meters and other devices view several times a day. The 1.5watt LED ligth is sufficient light, the lowest wattage CFL I have seen is 9 watts. Does it produce more light, maybe, but at 7 times as much energy. And, as has been discussed, the CFL life is reduced if turned on/off frequently. With a 1.5 watt LED there's not reason to turn it off anyway....not sure its life is impacted by on/off cycles.

In cold weather I also have some "standing" lights I leave on outside. I'd do it in warm weather too, but the bugs/insects are too much. In those application I will also try the 1.5 watt LED and see how it works. As my goal is to have enough light to see is something is going on in the near area, I think it will work fine, and at the reduced power consumption as compared to a low output CFL.

 

[DiscoInferno]

I assume that the 1.5W LED light is more of a directional (spotlight-type) light? Is it sufficient light everywhere, or just directly in front of the light? CFLs are pretty omnidirectional unless equipped with a reflector, so if you compare based on peak brightness (candelas) rather than total light power (lumens) the directional LED light will win. In applications that call for small spot lighting, this suggests using LEDs. When trying to light a whole room, this suggests CFLs (for now). What are the specs on the LED light?

I hunted around the internet out of curiosity, and I didn't find any commercial LED lights that did significantly better than the numbers given in the DOE document. A typical one I found - 2.3W producing about 100 lumens, which corresponds roughly to a 15-20W incandescent light bulb. What I didn't find is much in the way of 60-75-100W incandescent replacements.

As an aside, government career scientists are generally quite competent. Management, especially at the high levels where the positions are political appointments, not so much.

 

[Jerry_NJ]

The LED I'm testing, purchased at Walmart, is boldly listed on the package: Replaces 40 Watt, Use only 1.5 Watt. 90% More efficient than incadescent and halogen bulbs, 30,000 hour life. Mercury Free (good), and somewhat less eye catching "decor/accent" which suggests not for use as general lighting. My observation is the light is not a spot light, but appears to have a more central/downward brightness, and it is less light to my eye than one sees with a 40 Watt incandescent, or a 40 Watt replacement (9 Watt or so) CFL. The package does not give a lumens number as do the packages on incandescent (required by law I think). Again, to my eye this 1.5 Watt (I have not yet measured the power input) LED puts out less light than a 40 Watt incadescent, but clearly more than a 20 Watt....I must have one of these around and I give it the "eye test" to see how much light the 20 Watt puts out.

It may be like with flashlights. If I want to see over a distance of 200 feet or more outside I'll take my 14.4vdc lantern outside. I don't have any "big" LED flashlights, but I can say one of my 3 AAA cell 4 LED flashlights is a real wimp by comparison, but inside where I don't know that long distance light the 4 LED flashlight does just fine, and runs a lot longer on 3 AAA cells than does the drill battery driven hand lantern. One thing is clear to me, LEDs give usable light a far lower power consumption than any other electric lighting device I have experience with.

 

[jebatty]

If anyone with a "40 watt" or so equivalent LED lamp has a Kill-a-watt meter, please test amps, watts and PF for the LED and an "watt equivalent" CFL. That should make an apples to apples comparison with real data. I have the Kill-a-watt meter but no LED with which to do a comparison.

 

[Jerry_NJ]

Well I have both an A/C volt meter, and an A/C clamp on Amp meter with good low current sensitivity (multi-loops). So I can easily measure volt-amps and the Watts used is less-than or equal-to VA. In the subject case I expect to measure an A/C current less than 10 ma. I'll "publish", after I watch the Palin/Biden discussion. Maybe not until tomorrow.

 

[jebatty]

otherpower.com reports the following info on lighting efficiency. White LED's according to this article are about as efficient as incandescent. T8's look the best way to go for shear efficiency.

* 32 watt T8 fluorescent--85 to 95 lumens/watt
* standard F40T12 cool white fluorescent--60-65 lumens/watt
* compact fluorescents--low 30's to low 60's lumens per watt, usually 48-60
* UPDATE 12-2006 -- The newest white LEDs shipping from manufacturers in 2006 are approaching the efficiency of compact fluorescents, into the 50-60 lumens/watt range. At this time, however, there are no RE lighting products that use these new LEDs, and they remain extremely expensive. Manufacturer's research labs are also reporting new white LEDs that approach the 80+ lumens/watt of T8 fluorescents. However, these products are not shipping yet.
* T3 tubular halogen--20 lumens/watt
* white LED--15-19 lumens/watt
* standard 100 watt incandescent--17 lumens/watt
* incandescent night light bulb (7w)--6 lumens/watt
* incandescent flashlight bulbs--dismal, less than 6 lumens/watt

http://otherpower.com/otherpower_lighting.html

 

[Jerry_NJ]

Thanks "jebatty" this is the kind of data, even if a surprise to me, that I've been looking for, apparently not hard enough. I had posted questions about LED applications for residential lighting on another forum, may have been either electronic or electric, and I didn't get much. I'm not sure why different fluorescent "geometries" have different efficiencies. I'll check, but I believe the T8 is a circular bulb, and it has significantly higher efficiency. I had moved almost totally to CFL, but it is possible to use the round bulbs in my unfinished basement and garage applications just as easily as the CFL.


I'll do a cut-and-past into my Palm notes your good effort above, along with the url reference.

Still, I see some high efficiency uses for LEDs in my case. That's not to say I couldn't make equal or better use of a CFL if there was one using only 2 watts or less. The lowest one I have is a 40 Watt equivalent and 9 W.

More on the eyeball test: in my basement at night I see the 1.5 Watt LED a bit less light than a 25 Watt Incandescent, and a 9 Watt (40 Watt equivalent) has a bit more than the 25 Watt, but not a true 40 Watt replacement. So, I still see applications for the 1.5 Watt LED fixture I got from Walmart, some very inexpensive high duty (on) cycle applications.

 

[jebatty]

I think T8's are pretty much standard, including the 4' shop style lights.

 

[billb3]

A fair indicator of the lack of efficiency of some of the "super white" LEDs is the whopping size of the heat sink needed.

 

[Jerry_NJ]

Some home empirical results. The main problem with the LED measurements is they are too low for the electric power equipment test equipment I have. The lowest range on my clamp on amp meter is 6 amps. I used a loop of 4 turns which made the lowest range 1.5 amps full scale, thus I could see 100 ma, but not 10 ma.

I tested as a reference a 25 Watt Incandescent bulb (should be about 200 ma) to note the meter deflection. I then used a 9 Watt CFL specified at 140 ma (which suggests a large power factor if the real power is only 9 Watts) and noted the meter deflection compared to the 25 Watt Incandescent. In any case when I plugged in the 1.5 Watt LED (specified at 30 ma) I couldn't see a measurable deflection on the meter, suggesting it was less than the 30 ma. But, even if it were 30 ma the maximum power that could be consumed (all real power) is 3.6 Watts. So what do I conclude: I conclude the LEDs in the current crop of Walmart sold LED fixtures have a higher efficiency than do CFL. The 1.5 Watt LED fixture puts light over an area, not a spot light, that is at least half that of a 9 Watt CFL, or said another why 6 of these LEDs would swamp the light put out by a 9 Watt consumed CFL while using approximately the same real power.

 

[jebatty]

CFL's approximately use 1/4 to 1/5 the watts of an incandescent lamp for the same lumens of output. An incandescent is a pure resistance load with a PF of 1. Using my Kill-a-watt meter, I tested several CFLs, and the PF ranged between a low of .44 to a high of .7, making their effective advantage over incandescent of power used (VA) for lumen of output ranging between a little less than 2 to about 3. The newer CFLs had much higher PF than the older CFLs. All the CFLs I tested had electronic ballasts.

I believe most homes are charged for their electricity based on watts, not PF (contrary to many businesses). So cost-wise, CFLs do have their advertised advantage.

 

[Jerry_NJ]

Interesting, a power factor as low as 0.44 on some CFL would cause a simple amp reading to be much higher than expected watts/volts, which I think I saw in my measurement. Not sure what the case is for LED fixtures, but I can't imagine there is any significant reactive power in them, i.e., power factor should be 1. In any case, I was happy to have difficulty even detecting the A/C current into my 40 Watt (claimed 1.5 Watt consumed) LED fixture. Thus I'm willing to estimate that the fixture is consuming the 1.5 Watts, not some higher number due to losses not included in the advertisement/claim. Again, the LED fixture does have 30 ma printed on the bulb base, suggesting it can using either twice the 1.5 Watts or there and a significant "imaginary" part (power-factor < 1).

 

[jebatty]

I think I'm correct in stating that an LED will "waste" at least 1/2 the power delivered from an AC circuit. And probably a little more if a rectifier diode is used along with a resistor/and or a capacitor.

I have a 120v AC LED night light, 3 white LED's, uses a bridge rectifier with a resistor. Put in on the Kill-a-watt: 1 watt, 7 VA, PF of .22. But very interesting is that readings stay the same whether or not the LED is "on." It is consuming power all the time, not just when it is "on."

 

[Jerry_NJ]

Thanks, not sure anyone else is reading, a PF of only 0.22, due to the capacitor it seems. I'm surprised the design went to the trouble to add a capacitor, across the output of the bridge rectifier to act as a (HF) filter, i.e., to make the rectified ac more dc-like.

I'm not sure what you mean about the LED being "on" verses "off". That is in my case the on/off switch is in the ac feed line, so the current/power is zero whenever off. If the fixture has its own switch and that switch is on the LED side of the "power supply", then the capacitor could sit as a load even when the LED is "off". The good news is this is reactive power, i.e., imaginary, and you will not be billed by the power company for that borrowing of power. In fact, power companies like capacitive loads, they help cancel the local inductive (motors) nature of most power loads.

The only LED fixture I have is sealed and everything fits into a clearly visible internal plastic-like structure that could easily hole a full-wave bridge and capacitor, even a large value capacitor as is needed to deal with low frequencyies, 120 pps on the output of the bridge.

I'm a EE, albeit it a manager for the last 30 years, so the subject is interesting to me and testing my long term memory :red:

 

[billb3]

I think I'm correct in stating that an LED will "waste" at least 1/2 the power delivered from an AC circuit. And probably a little more if a rectifier diode is used along with a resistor/and or a capacitor.

I have a 120v AC LED night light, 3 white LED's, uses a bridge rectifier with a resistor. Put in on the Kill-a-watt: 1 watt, 7 VA, PF of .22. But very interesting is that readings stay the same whether or not the LED is "on." It is consuming power all the time, not just when it is "on."

If you open the AC circuit, there's no AC power to "consume".

 

[Redox]

The only thing it is wasting is time.

The LED is only going to conduct in the forward direction. On the negative half cycle, it actually goes dark and is really flashing 60 times a second and will appear dimmer to the human eye (that whole persistance of vision thing). The full wave bridge is converting the negative cycle to a positive and this makes it look brighter with the same current. It's still flashing 120 times a second, but the dark phase is very short.

As for the power factor, I would expect it to be 1 as the major part of the load is the dropping resistor, but I have no way of testing this. The problem is that the Kill-a-watt is probably not going to be very accurate at such a low load. I have compared the K-a-w with a Fluke power analyzer and it appears to be very accurate with loads of several hundred watts, but at the single digit level, I can't tell. I have plugged a few CFLs into it and they do seem to have a poor power factor, but this doesn't matter with a typical residential meter.

I have only seen one LED lamp with a medium base that will replace a standard light bulb. Anyone know of any? I think they would be great in the refrigerator and maybe the bathroom. CFL's don't seem to last in frequent on-off applications for me.

Chris

 

[Jerry_NJ]

redox, the LED fixture I keep referring to has a medium (standard light bulb) base and I purchased it at Walmart. In fact this it the third purchase at Walmart, the other two went back for a refund when they failed in a few days. Both had been in a basement stairway overhead and were switched on/off 10-20 times per day. The one I'm now using, testing, is in a dark corner of the basement, and I turn it on/off once a day. So far so good.

As for the flashing due to the rectified sinusoidal ac waveform, that's what the subject capacitor is all about, could sit behind the bridge, but before the "power supply" output current limiting resistor. This capacitor will reduce the LED "flashing". It also adds some reactance to the load, thus a power factor less than 1.

I think I'll make a 10 turn "amplifier" for my clamp on amp meter so that my low range will be 0.6 amps rather than the provided 6 amps. This will allow me to measure ac currents as low as 100 ma with some accuracy. Of course this will not address the need to measure true power. I do have a dual trace oscilloscope, so I should be able to measure the phase angle between the voltage and current.

 

[DiscoInferno]

You're making me want to get out my Fluke Scopemeter and start playing with circuits again. It's probably been 10 years since I've done much on the hardware side where I wasn't mostly looking over someones shoulder. I'm a theory/math/matlab type EE nowadays.

Re: LED flashing - Some LEDs are actually capable of higher total light output if fed a high-current, low duty-cycle pulsed waveform rather than lower-current DC. Looks like that mainly applies to low-power LEDs, however, and not the white LEDs we're talking about. Basically they are more efficient at higher currents than the max steady-state current, but the duty cycle has to be reduced to keep them from burning out at the high current.

 

[Redox]

I'll have to check out that fixture, but I have an aversion to going in a Wal-Mart.

Why not just insert an ammeter in the circuit to see what is going on? It's going to be a lot more accurate at the current levels we are talking about than a clamp on. Meter accuracy depends on the range of the instrument and I suspect that your 6 amp analog Amprobe (really a 200 amp meter with a low range, IIRC) isn't going to be all that accurate when you can barely get the meter to deflect.

We have a few strings of LED X-mas lights that I find irritating because of the flashing. I suspect that they have no external rectifier and are just allowing the AC to be rectified by the LED. We also have some LED night lights that don't flicker bad enough for me to notice. I doubt they bother with a capacitor, but a bridge might be possible.

I was under the impression that LED life was determined by the peak current and solid state just doesn't tolerate overcurrent well. The current, intensity and lifespan are going to be determined by the resistor you choose. I recall a SMT pick and place machine where I used to work that had a vision system with LEDs that only lasted about 20 hours due to the high current. You can push any LED closer to the edge, but will sacrifice their lifespan. I'm getting kinda used to not changing the CFLs as often and look forward to a lamp that lasts for 50000 hours. Now if we could just make it bright and white...

My new TV set has an LED light engine. No problem with brightness there!

Chris

 

[Jerry_NJ]

redox, yes on the amp meter sensitivity, that's why I mention a ten loop "amplifier". I think the "adding" of the magnetic flux is accurate, i.e., it will increase the sensitivity of the clamp on by a factor of 10. As for a meter in series, well then I have to break one leg open, and that then introduces the resistance of the series inserted meter, but that may not be a problem as the circuit has high impedance/resistance. I'll have to look to see what I have ... yes my Radio Shack Micronta digital multi meter, must be ten years old, does have an AC ma capability, so I do have equipment to directly measure low current AC. I may get around to that tonight, and think it may be easier to "insert" the meter than to make the ten loop magnetic amplifier.

 

[Jerry_NJ]

Spotted today in a Harbor Freight sale paper (mailed to my home, boy is that old fashioned?) an A/C power meter, for only $30. It is hard to see and has very little information in the "flier" but it does state limits of 120 vac and 15 amp ac, reads in Kilowatt hours and measures power factor. It has a digital display. The ad doesn't say anything about the ranges over which it can be used, but with a digital meter it may be capable of displaying small values, e.g., watt-hours or smaller. Here I assume the display in in KWH per hour. I may buy it.

 

[Redox]

It it this one?

http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=93519

I think that is the older model and it loses the info if the power goes out, but it seems to be accurate with a couple hundred watts to measure. I think a meter that measures 1800 watts is going to have a problem with a fraction of a watt. Remember that meter accuracy is measured as a percentage of full scale reading and just because it's digital doesn't make it any more accurate.

Chris

 

[Jerry_NJ]

Thanks Chris, I tried to get some specifics on sensitivity, and found some better close-ups of the meter than HF gives. In fact "smarthome.com" offers the same meter for a few dollars less.

I understand the 0.2% accuracy applies to full scale, but it specs some impressive accuracy. That's why I was trying to find out if it is possible to change the "scale". It appears it from the front panel that there are no range settings. It does show one button as KWH/Hour, so maybe that means it can be used to time something or to show KWH (per hour, I assume).

I may purchase, it looks like nice "toy" for everything but LED level units.