# Technique to measure/infer Vf for LED given If

This isn't the normal resistor question, and I'm familiar with the theory and Which is better: using a resistor with an LED, or "precisely" matching the Vs to the given LED's Vf rating?

I have one of these impressive LEDs. I'm intending to build a lamp out of it. The max forward current is listed at 720ma. The Vf is given as a broad range of 12-14V, which is fairly useless. On briefly applying a 12.0V supply to it through a multimeter I observe it consuming about 850ma. Clearly I need a small resistor in there, but is there a sensible way to work out the value required other than stacking increasing numbers of 0.1 Ohm resistors in series with it and plotting my own Vf/If graph?

(The device seems to be 12x strings of 4 individual dies wirebonded to a thin PCB on an aluminium bar, where the dies are covered in translucent material to act as protection and diffuser.)

Is a Vf/If graph likely to be stable as the device heats up?

I have an IR noncontact thermometer. Is this likely to give sensible readings when pointed at an LED, or will it over-read?

## 2 Answers

For such a high power LED, I'd suggest using a constant-current driver of some sort rather than relying on a resistor. With a low value resistor, you have the potential for a lot of variation in current, as you observe, and with a higher value resistor you need a lot of voltage overhead and waste a lot of power.

Besides dedicated LED driver ICs, you can construct a constant current supply quite simply using an opamp, a transistor, and a shunt resistor to measure the current.

• With the adhoc current limiter, suppose I use a 0.1R low side shunt; that gives 70mV. Does that work as an opamp input or is it too close to the ground rail? Edit: I guess the answer is yes, given electronics.stackexchange.com/questions/144285/… Feb 25 '15 at 13:33
• That depends entirely on the opamp; there are plenty of cheap opamps that can sense down to (and even below) their negative rails, though. For such an opamp 0.1 ohms - or even lower, potentially - would be just fine. Feb 25 '15 at 14:10
• What's the phrase to look for in a datasheet that describes that property? (Do you have a preferred example?) - thanks Feb 25 '15 at 14:34
• "Input voltage swing" is the relevant phrase. MCP6002 is an example of an affordable opamp whose input includes the negative rail; I use it on a constant current load I produce for hobbyists, in fact. Feb 25 '15 at 14:41
• Ah, re:load is yours? Nice. MCP6002 supply voltage is only 6V though, whereas the traditional LM358 can handle 12V and claims the ability to operate at the negative rail as a bullet point on the datasheet, so I might try one of those. Feb 25 '15 at 15:23

Assuming an LED or string of LEDs with no current limiting built in:

The simplest way is to connect a constant current supply which is set to 700mA and can provide at least 14V, preferably more. Note, that's a constant current not a current limited supply.

Then measure the voltage across the LED. That will be the forward voltage.