No, this is a constant current buck supply, with a series inductor to maintain nearly constant LED current. I think the current is held within 10% of the desired current. The reference design with a long evaluation report is almost exactly matching what I want to do, and they show a 94% efficiency at 350 mA. The LEDs just see DC.
Jon
We already have those in the kitchen ceiling.
Jon
LED tape is cool. You can buy it by the reel, and put in as many strips as you need.
--
John Larkin Highland Technology Inc
Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators
The consumer-grade LED stuff is awful, using panel indicator grade LEDs. They will keep you from bumping into walls at night, but at least the stuff I've seen can't light up a room.
10 of these LEDs will light a room quite well.Jon
Well, I got the LM3404HV reference design built and it does as advertised. I haven't done detailed measurements, but the finger test says the losses in the chip, inductor and Schottky diode are QUITE low, maybe just a couple degree C temp rise on the inductor, no rise detectable on the semis!
I think my estimate that this bank of 10 LEDs is brighter than a 36" T12 fluorescent has to be wrong. The fluorescent tube spreads light out over a very wide area (it has a white reflector behind the tube). The LEDs have domed lenses on them, and project a lot of light forward and much less to the side. This could actually be quite good for room illumination from a ceiling fixture, but it makes an already difficult comparison pretty much impossible. So, while it lights up a spot directly where the LEDs are pointed very well, it doesn't illuminate the whole room like the fluorescent tube. I'll get this thing jury-rigged in the kitchen and see what it looks like.
Jon
e .
A very crude but decently effective informal way to compare the total output of different sources is to A/B test them, pointing them at the ceiling in a darkened room. You, meanwhile, look at the floor, or objects surrounding you.
It's a rude integrating shpere[sic].
It'll be harder to do stopping and starting a fluorescent fixture-- maybe you could shutter it in and out?--but maybe the principle will inspire you.
I've used it extensively to compare LEDs with various outputs and patterns. Otherwise, if one's laser-focused and the other's Lambertian, it's hard to eyeball who's really making more light.
-- Cheers, James Arthur
Yes, that was what caused me to make a VERY optimistic judgment the first time I tried the LEDs over my workbench. I still think these things are going to work OK, but I will probably go to about 15 LEDs per panel and will need two of them to replace a dual
48" T12 fixture. That will be about 30 W input to replace a 60-80 W fixture, and get longer lamp life plus instant start without problems when humidity gets higher. The old fixtures won't start reliably in the summer.Jon
Le Mon, 01 Apr 2013 11:52:52 -0500, Jon Elson a écrit:
I'm rebuilding my bench and look for good lighting. I had a look but too many refs. Those surely aren't the $10 or so LEDs. Any part number?
-- Thanks, Fred.
Your assessment may not have been entirely off--focusing light where you want it is a big advantage, and lots easier with LEDs.
I rigged a 1 watter from the ceiling over my bench this weekend. Tightly-focused, it makes a bright 5" "inspection" area, mostly replacing a 50W halogen flood.
-- Cheers, James Arthur
? 2013?3?8????UTC+8 ??12?49?56??Jeff Liebermann ???
ot
....
There are some now appearing that are in the 80Lumen/W class and I was caught out by a 10W nominal 60W equivalent one that was too bright for the use on my parents stairs (where a CFL took too long to warm up).
The colour temperature claimed was "warm white" 2700K but it was actually closer to 4000K - apart from that it was a perfectly good clean light. The heat sinking means that the upwards light emission is much less than with a conventional bulb envelope but it was a first for me. An energy saving lamp that was actually "too bright"!
Too early to say if they cook their PSU capacitors to death the same way that CFLs do. First impressions are it runs cooler.
-- Regards, Martin Brown
XPEBWT-01-0000-00CC2CT These are $1.73 in single quantity at Digi-Key, I'm running them at about 300 mA right now using the LM3404HV. You have to buy 50 to get the first discount step. They are rated at 104 lm at 350 mA, and that gives 102 lm/W. Scanning a whole bunch of pages, this was the best lm/$ part I could find about a month ago. I took a piece of blank PCB material and scribed lines breaking it up into squares of about
1.5 sq inch each, and then soldered the LEDs across the gaps. I think the LEDs themselves run about 50C this way, maybe less as my fingers get heated by the optical output.Jon
The Cree XPEBWT-01-0000-00CC2CT available from Digi-Key for $1.72 EACH is rated at 102 lm/W (not counting power supply losses, of course). You sure can't look directly at the LEDs!
Jon
Making LED lamps is really simple. A simple power supply, say with 1A current output can easily power 3 - 4 LED lamps, each with say 30 LEDs. For intense light, please connect yje LEDs in parallel, each with a dropping resistor. There are are some designs though, which exploit capacitive reactance, but the current output is low, and so the LEDs need to be connected in series, and consequently the light output is low. More importantly, there is no galvanic isolation, and so the design is inherently dangerous.
That 104 lm/W is measured at 350 mA (1 W), so you need 7 of those for a 60 W equivalent. At such low current, the huge life time claims are believable.
The maximum current is 1000 mA, but the efficiency and lifetime drops drastically at such high current.
On Wed, 03 Apr 2013 17:36:45 -0500, Jon Elson wrote as underneath my scribble :
That looks very efficient, any figure on the total consumption ie as seen from the mains including powersupply losses? My guess you will be closer to the commercial cree based bulbs then. C+
When doing efficiency comparisons, one should remember that traditional fluorescent lamps and other discharge lamps (such as mercury), the power refers to the power consumed by the lamp itself. In addition to this, you must also include the losses in the ballast/electronics.
OTOH, E14/E27 base CFLs are rated according to the total load.
Looks like about 7 W DC input, but then there's a transformer and rectifier. I think mabe the isolation of the transformer is a good idea in this experimental stage. Later, maybe an off-line supply would be more efficient.
Jon
Right! And, my guess, from the temperature, is that the usual magnetic ballasts are QUITE inefficient, got to be burning about 20 W for a dual
48" T12 ballast. The newer electronic ballasts are more efficient, but I don't know how much. I know the transformer I used for my prototype is pretty efficient, maybe 1-2 W loss at the most.Jon
I've run my string of 10 up to about 400 mA while testing, but plan to run at either 300 or 350 mA in the final setup. I may decide to run a string of 15, which is about the limit of the LM3404HV chip. I'm comparing to a dual 48" T12 fluorescent fixture, and the 10 LEDs at 300 mA doesn't come close, when tested in our kitchen. Our laundry room/pantry is a real dark nook, so maybe this prototype will end up there in a couple of days.
Jon
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.