Procedure to getting Voltage and current of an LED without knowing ANYTHING else?

Good day good people. I got batches upon batches of various surface mount and dip leds. Not a single god damn seller wrote the specs on neither the website, nor the package and there are so many that I can't tell which one's from who. Regardless, however the ridiculously low price you get these things for, you can't really complain.

So I need to know if there is a procedure to finding out the current and voltage of an led and that is without knowing a single rating about it. It it guesswork or is there a proper testing procedure? It can be a 1,5v can be a 2.5 or a 3v led I don't know. What should be tested first, what should be second?

All I have is a basic multi-meter and besides checking batteries voltage I have't been using it for anything else.

Thank you, folks.

sort by: active | newest | oldest
Yonatan242 months ago

Well, if you have a lot to ruin, why not do some testing? :)

Personally, I've tried giving 2.85V to LED's that I found, thinking that it would work at 3V. 1.5 years later, and they still work like new.

Don't tell the EE's, they'll kill me!

iceng Yonatan242 months ago

I am smiling that you are learning some electronics and hope you use a series resistor when you light up LEDs today...

A beginner might misconstrue comments and forget to use a series resistor and Ruin an LED...

Yonatan24 iceng2 months ago

Not today, since I have evidence that I don't need it! Well, at least for the applications that I use my 10W 12V LED's for.

I have run ultra brights from Radio Shack on 3 volt button batteries and no resistor.

I've been told that those 2032 coin cells aren't able to supply enough current for burning up an LED.

I connect my 12V 10W LED's in parallel to a high-speed dimmer, and to a 12V power supply. I make sure to keep them not more than warm:




None have ever failed on me so far. None, either than the ones that have failed during my dreams after having tons of EE's comment that they will last an hour if I'm lucky... :)

I'm an EE and I argue with EE's all the time. Too many of them think 0 tolerance or accurate to .0000001.

You build it and it works that is all the proof you need.

All electronics are ballparks; stay in the ballpark and it works, knock it out of the ballpark and it's game over.


I'm always specifying the +/-limits with the understanding that the smaller the tolerance the more it costs to make.

I worked with a lot of components with 20% tolerance.

The difference today no one wants to calibrate.


True why calibrate when a uP is available... My reference to tolerance was in say 4" by 3.5" PCB. Some drafter might specify 4.000" by 3.500" when a simple 4.0" x 3.5" +-o.05 is plenty including mounting holes referenced to a datum hole and a good deal less expensive to fabricate....

If he was using CNC I can see the CNC needing that; but he can't convert from 4 to 4.000 and 3.5 to 3.500?

I had a buddy that made parts for Arospace.

I asked him to make a part for a antique gun and he couldn't make it because it was not a tolerance of 1/10 of 1/1000 of an inch. By the way depending on the part of a gun they have tolerances up to 20/1000 of an inch and more on the shell casings.

As I have UV LEDs across the same batteries and no resistor....

But these are special cases and a newbie LED user should not assume LEDs are used without some form of current limiting....

BTW a brand new CR2032 coin cell delivers a half amp short circuit that falls to about 300ma in half a second and 160ma after a second...

Some pictures of UV, Red, Yellow and clamp up close. Then there is a good example of 125 LEDs all wired in parallel (notice some are dim and bright because there are no series resistors)

Yes, the internal coin cell has enough current limiting to work LEDs without the obligatory series resistor...

Do click the pic to see the whole image....


I like your battery holder.

On average they work on 5 to 20 ma.

10 ma should be fine.

However if you want to have fun take 220 ohm resistor, a 5 k ohm pot a 12 volt supply.

Connect the pot like the volume control in this circuit with 12 volts on the input.


I have a number of Instructables on Reverse Engineering.

This circuit will allow 0 to 55 mA current through the LED.

Connect the LED and the meter from the pot adjust to ground and adjust the pot starting at 0 volts up.

If the LED doesn't light up turn it around you are reversed biased.

When the LED lights up enough measure the voltage across the LED with your meter and that is the forward voltage.

Then measure the current through the LED and you have the current.

Answer 142.bmp


iceng2 months ago

Please regard a 1206 SMD LED forward voltage Color Chart by clicking it to view the entire image.


Nice chart I think I will save that Image.

Downunder35m2 months ago

Apart from the good suggestion already given I had the same problem with a mixed batch of 250 LED's.
Not only did they come all loose and in one single bag, but also with no details on them at all.
First order of business was using the diode tester on the multimeter to sort them by colors, they only glow a bit but enough to see what color they are.
Also helps to find right polarity for unknown LED's ;)
Once they were sorted by size, type and color I took three out of every batch for a testrun.
I used the standard voltages as stated in those many LED calculators.
But for each color and type I selected a different resistor to change the mA accordingly.
20mA is a good starting point for most LED's, although SMD types and red LED's are often happy just 15mA.
So I basically got the first LED of a series running on the standard voltage and amps as the clculator suggested.
The second 10mA more and the third another 15mA more.
Left them on continously for 2 weeks.
Some of LED's running on high current changed color or burnt out, so I knew the limit was set to high.
For the bigger ones showing no problems at all I did another run with even more current.
Especially newer blue and white LED's can sometimes run on 2.5V @35 or more mA.

Another test that can be done is a temperature check but this is a bit more complex.
Using some tiny clamps, heat transfer paste and a tiny temp sensor you can check if the temp of LED increases.
Normal LED's won't get warm at all, high power white and blue often do a bit.
So if the temp increases within the first 15 minutes of running to more than 10° above amient temp the LED is running to high.

Either way be prepared to kill a few of your LED's if you want to find the upper limits.
If you want to play it safe use the standard voltages and increase the mAmps until the brightness is sufficient.
Especially super birght LED's will thank you with endless years of running time.

Jack A Lopez2 months ago

Well, the complete picture, is an I-V curve,


However, the I-V curve for every LED is essentially the same. It is sort of hockey-stick shaped, so much so that it is actually a reasonable to approximate that curve with a straight vertical line, essentially a forward voltage that is constant, or almost constant, over a wide range of current.

As an example, a typical I-V curve for a LED, might have these three points on it:

V= I=

2.3 V 0.1 mA

2.4 V 1.0 mA

2.5 V 10 mA

The voltage does not change much, a few tenths of a volt, even though the current has changed by by a factor of 100, over the range from 0.1 to 10 mA.

In fact you could probably imagine the curve continuing this way, current increasing by a factor of 10, for each 0.1 volt increase in voltage, e.g.

V= I=

2.6 V 100 mA = 0.1 A

2.7 V 1000 mA = 1.0 A

2.8 V 10000 mA = 10 A

However, this part of the curve is not safe to explore in a real circuit, because these currents getting enormous, more than enough current to burn out the LED.

So, the usual approach is to pick an upper limit for your current, and then solemnly vow, "When I test this batch of bargain LEDs, I will not let the current through any in-circuit LED exceed the limit Imax."

What is good choice for Imax? I think Imax= 5 mA is a good number. That is probably about half, or a quarter, of the true maximum current for little discrete LEDs, like the kind you describe. However, it is certainly enough current to make the LED emit light you can see, provided it is a visible light ED.

Next you build a circuit for to give a single LED some current, not exceeding Imax.

As sort of an easy example, consider a 5.0 volt DC supply, wired in series with a 1K resistor, wired in series with a single LED.

The maximum current such a circuit can supply (i.e. the case where the LED is replaced with a piece of wire, or a paperclip) is just (5.0 V)/( 1000 ohm) = 5/1000 A = 5 mA

Supposing, as an example, you put a LED, with 2.4 V characteristic forward voltage, in this same circuit. Then the 1 K resistor has 2.6 V across it, which means the current is (2.6 V)/( 1000 ohm) = 2.6/1000 A = 2.6 mA, which, predictably, is a number less than 5 mA.

This simple circuit, combined with a voltmeter, gives you a way to measure exactly one point on the I-V curve of whatever LED you plug into it.

If you want, for the sake of clarity, I can draw you a picture, circuit diagram, of this circuit.

For a slightly more complicated LED tester, you could use an actual constant current source, set to your Imax = 5 mA, or whatever number you choose. This circuit also gives you exactly one point on the I-V curve, but that number is guaranteed to be:

(Imax, Vforward)

Upon request, I could draw a diagram of that too. The one I have in mind is the constant current source made using the LM317 voltage regulator IC, and you could probably find that one via Google Images, or searching for it here at Instructables, ( I know I've seen it here before.) faster than I could draw it.

iceng2 months ago

Assume the lowest voltage a red LED and plan on a 15ma current.

Now I will use a 9v battery example, but you can plug in 6v, 4.5v or even 12v.

The led MUST have a series resistor here is how to calculate it

9v battery-1.5v led = 7.5v across the resistor

Resistor = Voltage / current = 7.5v /o.015A = 500 ohms

It will be less on different color led and lower voltages.

Other cases R will be more....