Fixed threshold LEDs ?

Hi,

LEDs have differing forward voltages relative to their colour - increasing from IR to deep blue over a range from under 1V over 3V.

That is, unless they are intended to be party lights.

Multi coloured ( red, amber, green and blue) ones I have seen lately have a fixed voltage of 2.90v at 5mA.

Such LEDs can be run in parallel with similar light output for each colour.

All are high efficiency too with usable output at under 100uA at 2.5V.

What is the trick being pulled ??

FFS do not try to tell me it's an inbuilt resistor.

..... Phil

Reply to
Phil Allison
Loading thread data ...

They are inherently blue LEDs as in "white" LEDs, i.e. they use a fluorescent material to generate desired colour. Using a spectrum analyzer shows that the colour peak is quite broad.

True red, green and blue LEDs have a much narrower spectrum and different forward voltages.

Reply to
upsidedown

-------------------------------

** You are 100% correct - many thanks.

Under magnification, I can see the coloured phosphor on the red, amber and green LEDs but nothing on the blue ones.

BTW you can get a 100 of them in four colours from Wollies etc for about $15 with SMPS to boot.

Come with long attached leads and make ideal indicator lamps on mains and battery powered items. Fit tightly in a 5mm hole with a drop of super glue.

Funny they don't seem to be available from electronics dealers ?.

..... Phil

Reply to
Phil Allison

Hi. Can you name a few. Those kind of LEDs could be useful sometimes.

Regards LM

Reply to
LM

In China they probably match the forward voltages of the LEDs in a given string by hand-sorting/binning the raw LEDs; the manufacturer might be convinced to match a bunch of LEDs to whatever close forward-voltage tolerance you like for parallel operation but probably have to order in quantities of a few hundred thousand to motivate them.

Reply to
bitrex

easing from IR to deep blue over a range from under 1V over 3V.

have a fixed voltage of 2.90v at 5mA.

olour.

.

and green LEDs but nothing on the blue ones.

t $15 with SMPS to boot.

nd battery powered items. Fit tightly in a 5mm hole with a drop of super gl ue.

That sounds like an unfounded load of... conjecture.

Even in China, other than back alleys where people hand disassemble parts f rom PCBs under tarps, labor has far too much cost to fuss with manually sor ting LEDs. If they wanted to do this cheaply, why wouldn't it be done by m achine such as when they are being tested on the assembly line???

Because it's not worth the bother.

--
  Rick C. 

  - Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

easing from IR to deep blue over a range from under 1V over 3V.

have a fixed voltage of 2.90v at 5mA.

olour.

.

and green LEDs but nothing on the blue ones.

t $15 with SMPS to boot.

nd battery powered items. Fit tightly in a 5mm hole with a drop of super gl ue.

I doubt if they do it by hand. When I was working in the Netherlands (areou nd 2002) we wanted a LED that produced a fixed number of blue photons when driven by a given current.

I'd initially over-designed the circuit so that we could cope with the rang e of output intensity specified by the data sheet, but my boss wasn't enthu siastic, so we went to the UK manufacturer and asked how much they'd charge for a batch of 500 LEDS selected for a +/-5% brightness at particular curr ent.

It wasn't much. So I asked for a price on a batch of 100, which was only ti ny bit more, and that was what we bought.

The product was never that popular but they may have got through the batch by now.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

It assumes the lil fly-by-night novelty manufacturers in Shenzen that make the things knows who manufactured their LED supply in the first place.

String probably cost $2 AUD retail oh yeah I bet they're getting all well-specced batches of high-quality shit direct from Osram.

Reply to
bitrex

Oops, Phil said it cost $15. But they did include an SMPS that's nice.

Reply to
bitrex

ng from IR to deep blue over a range from under 1V over 3V.

a fixed voltage of 2.90v at 5mA.

r.

I sold over a million 5mm LED's to a client in NZ , factory direct with sin gle bin specs like +/-0.1V +/- 2nm and +/-250'K of course at premium with 5

0% margin. They were sourced from the best epiwafer with my specs and I ha d to spend weeks down there in the early days) hand-holding thru ESD issues until I got diodes put in everything.

I learnt a lot about process-controls and semiconductor physics. The indus try has improved their processing tools and the result is lower Rs (or ESR) and higher efficacy.

I came up with a rule of thumb for all diodes that was basically just like how all Zener Diodes are specified with Threshold voltage and Knee Voltage and Incremental Resistance, Rs at each level.

When you look up Vt @ 5% to 10% of the rated "If" (fwd current ) , we call this the approximate Threshold for safe operation and this eliminates the v oltage drop from electrode interface & bulk resistance, Rs.

You can make a linear regression estimate of

Vf= Vt(@5 * %If) +Rs*If at 5mA you are seeing about a 10% reduction in v oltage to due this Rs only. Vt does not change much with batches, only che mistry and wavelength because the activation level is proportional to an en ergy level which is due to the shorter wavelength.

Don't worry about 5% difference because the tolerance on Rs in the product used to be 75% and is now typically 25% which causes all the 25'C variance in LEDs. It reduces sharply with the rising cost of the Chip due to qualit y and power.

Here's the other Fact I learned about ALL diodes.

Rs ~ k/Pd the max DC power at 85'C, Pd rating for k= 0.3 to 1. 1 is typ ical and 0.3 for the best Korean power LED's.

So a 65 mW 5mW LED is about 16 Ohms for Blue, Green White but around 12 Ohm s for Red, Yellow. You can bank on that value with the threshold voltages if you can measure that with your DMM diode mode or use a resistor and 9V battery for 2~3mA. These thresholds for White are typically 2.8V and 3.1V for 20mA .

But in truth it is a logarithmic curve and this is just a simple linear reg ression estimate.

Reply to
Anthony Stewart

I would expect a party-light manufacturer is buying a mixture of stuff that's already been picked over at least once, for cheap.

If the Rs tolerance is that low by design nowadays sounds like they could just statistically sample whatever mix of stuff they have and derive a population estimate for the Vf that some large percentage of the LEDs they have to work with today fit in, and roll with that. some strings will be duds and they get rejected on the way out by QC but most work OK, so cheap to make to begin with no biggie if 5% or something of the finished product getting rejected.

Reply to
bitrex

If you are on about forward voltages, what are you after, controlling brightness ? I looked into that not long ago and determined if that is what you need use PWM. It is easier to control than even series strings on higher DC.

Sorry if I read the question wrong.

Reply to
jurb6006

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.