LED brightness

[peterg]

I'm looking for a chart or calculator that shows what value resistor to use with LEDs to obtain equal brightness regardless of the LED's specs. ie if a 4k7 resistor is used with a  4000 mcd red LED what value resistor is used with an 8000 mcd green LED to get the same brightness. I've done some searching but can't find anything straight forward. I've also done some trial and error testing with varying resistors but want to get more precise.

[deadastronaut]

i dont think you can beat setting them up on a breadboard with different resistor values really and going by eye..

not much help...but that's my 2p.

[armdnrdy]

I don't think you're going to find this. There is no exact consistency in LED brightness. If you are applying LEDs side by side and want them to match...you have to tweek the resistor value or use a trimmer.

I'm looking for a chart or calculator that shows what value resistor to use with LEDs to obtain equal brightness regardless of the LED's specs. ie if a 4k7 resistor is used with a  4000 mcd red LED what value resistor is used with an 8000 mcd green LED to get the same brightness. I've done some searching but can't find anything straight forward. I've also done some trial and error testing with varying resistors but want to get more precise.

[peterg]

Thanks Guys - I'll keep working with the trial and error method.

[davent]

I built the Mark Hammer LED Tester which is just a 1×12 rotary switch with a range of resistors to suss out what works best for any given LED/situation. Also can switch in another 10k resistor in series so 24 values to quickly switch between.




dave

[R.G.]

LED brightness has two components. One is how naturally bright the particular LED is per unit of current, and the other is how much current is being put through it.

The brightness per current is different for different types and colors of LED. LEDs put out light that is nearly linear with current. So an LED goes from no light at zero current up to its maximum brightness at its maximum current. The maximum current is generally determined by what make the LED too hot. For a standard T 1 3/4 style LED, the maximum DC current is usually about 20ma.

Some LEDs are so dim that you can barely see them in sunlight at 20ma, and some are downright blinding at 20ma. But if you turn down the current, you turn down the brightness.

The amount of current has to be limited by some external device. If you apply an increasing voltage across an LED, the LED will not conduct at all until it hits some minimum forward voltage. After that, current rises exponentially with any additional voltage.

So the normal way to run them is to apply a voltage bigger than the forward voltage, and then to put a resistor between the voltage source and the LED. The resistor sees a voltage equal to the difference between the voltage source and the LED voltage, and limits the current to that voltage difference divided by the resistor value.

So - pick an LED based on its color and milli-candela rating, then pick voltage sources and resistors to cut that light rating back down to how bright you want it to be.

[Gurner]

I built the Mark Hammer LED Tester which is just a 1×12 rotary switch with a range of resistors to suss out what works best for any given LED/situation. Also can switch in another 10k resistor in series so 24 values to quickly switch between.

or if you're lazy, just buy one of these cheap & cheerful units...

http://www.ebay.co.uk/itm/Mini-Handy-LED-Tester-Test-Box-2-150mA-for-Light-emitting-Diode-Bulb-Lamp-T1K-/380675626722?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item58a20916e2

establish what preferred current is best for your required light output (by placing the LED in its different breadboard connector type slots) & then it's a simple formula....

series limiting resistor required =  (your circuit's supply voltage - the fwd voltage of your LED) / your preferred current

So for example, you put your LED in that little unit & decide that 10mA is best for that LED. Now, in your deployed circuit, if your LED will be supplied by 9V & the fwd voltage of your LED is say 3V, therefore (9V-3V)/.01 = 600 ohms series limiting resitor.

LED brightness has two components. One is how naturally bright the particular LED is per unit of current, and the other is how much current is being put through it.

I would add to that the lens ....typically, LED brightness is inversely proportional to the overall viewing angle & the viewing angle is dictated by the LED lens. In other words the smaller the viewing angle, the brighter the LED looks...a lot of high brightness 3mm LEDS have a viewing angle of just 20 degrees & look intensely bright, whereas say a 1210 surface mount LED has a 120 degree viewing angle ...they look far less bright (but appear to have the same brightness over a far wider viewing angle).

To compound matters, different wavelengths appear different to the human eye ...therefore ( I think!) it's green that looks brighter to the eye than red, so even with equal current, equal spec LEDs, equal lenses....one colour LED will still 'look' brighter than the other!

[Mark Hammer]

That's a neat little gadget!  And Dave's adaptation of what I did is pretty cool too.

To RG's always-authentic description, I would add a 3rd and 4th dimension which is:
a) how sensitive the human eye is to that part of the spectrum (not all wavelengths are perceived with equal sensitivity), and
b) the degree of available contrast where the LED is mounted.

So red and green LEDs of different sizes, and set against different backgrounds, may requires different amounts of current to be seen at equal brightness than one would predict, given their luminance ratings.

That's part of what makes the empirical method of varying the current with known series-resistance values so useful.  You keep changing the resistance until it hits the perceived sweet spot, and that's the resistance value you use for the eventual circuit.

If a person desires a little more precision, and a broader range of values, consider using an 8 or 10 element dipswitch, wiring an assortment of resistor values in series, and use the dipswitches to shunt different combinations of resistors and achieve different series values.  Consider how many different combinations and permutations there would be for the following eight resistors: 10k, 6k8, 4k7, 4k7, 2k2, 1k, 1k, 470R.  They total 30.87k, but cover the range between 470R and that upper limit, with a lot of waystations within that range.  Heck, even a 4-switch unit with 10k, 4k7, 2k2, and 1k will cover a lot of ground (1k, 2k2, 3k2, 4k7, 5k7, 6k9, 7k9, 10k, 11k, 12k2, 13k2, 14k7, 16k9, 17k9).

[Mark Hammer]

Okay, proof of concept.  I whipped this up using 5 resistors and a 4-switch DIPswitch.  The 5 resistors are wired in series, directly on the pins of the dipswitch.  Various resistances are achieved by shunting any combination of resistors.  Since it would be possible to have NO current-limiting if one accidentally closed all switches, I wired a 5th resistor as a default, to avoid any LED frying.  The unit simply mounts on the 9v battery, via the snap, and you press the momentary switch to test the LED in the socket.

The 5 resistor values are 1k (default), 1k, 2k7, 4k7, and 10k.  Total series resistance is 19.4k, and possible series resistances are 1k, 2k, 3k7, 4k7, 5k7, 6k7, 8k4, 9k4, 11k, 12k, 13k7, 14k7, 15k7, 16k7, 17k4, 18k4, and 19k4, if not more possible values.  That should do a reasonably good job of covering older "garden variety" LEDs right up to superbights..  Flick the switches until you find the brightness you like, and figure out the nearest common value.

[Jdansti]

Very nice! I'm going to have to make one for myself! 

[armdnrdy]

peterg,

Now you have you choice of testers to make the first LED/resister/brightness chart!

[Gurner]

Okay, proof of concept.  I whipped this up using 5 resistors and a 4-switch DIPswitch. 

I wouldn't use a 9V battery...as your results will vary quite considerably, depending how far along the fade profile the battery is (ie how long it has been used & how much life there is in it).

For example a fresh 9V battery is about 9.4V, but a battery on its last legs is near 6.5V....that's a 50% variance, which will give you wildly different results!

You really need to use a regulated supply for such a tester, or change the design to be more like that of the one I linked to on Ebay earlier on (which yields a constant current no matter what the battery voltage is)

[duck_arse]

.... and, if you stick an ldr inside a 35mm film can, or similar dark place, you can then throw it over the led-on-test, read the ldr resistance, and get some "numbers". which will of course vary with colour, distance from led, etc, but numbers all the same.