Measuring power LEDs to match voltages

Thread Starter

taith

Joined Apr 1, 2017
28
I have about 100x3W SMD LEDs, often called LED star, but these are without circuit board used as radiator.
Every LED have different voltage drop, and I plan to match them all in series of 4, so 25 parallel of LEDs, each in series of 4, resulting with contraption that I can regulate with adjustable 12V power supply.
How should I measure these LEDs to match voltage drop between all 25 pairs?
Will low power measurement be enough? My multimeter can declare their voltage drop at 0,3mA
Or voltage drop would be different at 900mA? If so, measuring them in high current mode might prove difficult, especially without proper heatsink, voltage drop changes depending of temperature.

Alec_t

Joined Sep 17, 2013
12,802
Ideally, since voltage drop is current-dependent and you can't rely on random LEDs having identical voltage/current characteristics, you should measure them when passing the current you will eventually be using. Since you are going to need heatsinks anyway, your test rig should include them.
BTW, at 3V per LED (nominal only), 12V leaves no control headroom so is inadequate for driving a string of 4 LEDs in series, unless you intend to drive them well below their maximum output. 3 LEDs in series would work; or a higher supply voltage.

MrChips

Joined Oct 2, 2009
26,077
Your design premise is wrong.
As you have discovered, LED forward voltage is very variable.
Simply control the LED current and be done with it.

mvas

Joined Jun 19, 2017
539
I have about 100x3W SMD LEDs, often called LED star, but these are without circuit board used as radiator.
Every LED have different voltage drop, and I plan to match them all in series of 4, so 25 parallel of LEDs, each in series of 4, resulting with contraption that I can regulate with adjustable 12V power supply.
It is best, for each string of 4 LED's to have its own 900ma Current Source.
Typically, we do not connect LED strings in parallel

If you want to experiment with using a single 12 Volt P/S to adjust brightness of all 25 strings then ...
add appropriate ohm & wattage resistor in series with each string of 3 or 4 LED's, to limit the 12 Volt source to 900ma per string.
With 4 LED's, the resistor will dissipate +/- 1 watts.
With 3 LED's, the resistor will dissipate +/- 4 watts <= Now do you see why we use a Current Source ?

You will discover that a very small change in voltage makes huge change in brightness.
Which is another reason we do not typically use a Voltage Source for large LED designs.

Last edited:

BobTPH

Joined Jun 5, 2013
4,888
I'd go with 15V, 25 strings of 4 and 25 LM317s.

By the way, what are you going to do with this? It will produce a lot of light. Equivalent to about 2KW of incandescent bulbs.

Bob

oz93666

Joined Sep 7, 2010
737
The way I would do it is to select at random and wire together your 25 lines of four in series ....

Then connect to a variable power supply and slowly increase the voltage , one line of 4 will start to light first ... increase voltage and note which line is the last to light up .... switch off power and change 2 leds from the first to light string , swap them with 2 leds from the last to light string ...

Repeat many times ....it's a bit long winded , but by this method you should get closer to balance ...

Balance does not have to be perfect ...the eye is very sensitive to low light levels ....

These can probably run at about 10% max recommended current with no heat sink ... this might be more sensible than going to all the trouble and expense of heat sinking ... to run at full power you will need an exceptional heat sink ,or you will get very short life.

I would wire 5 or perhaps 6 in series and run with no resisters experiment to see what is posible , run direct from your 12V supply .... UNDERDRIVE (use low current) below 50%... even with heat sink

Last edited:
Similar threads