Why 3V LED still requires resistor when input voltage is also 3V

Thread Starter

Kelvin Lee

Joined Oct 22, 2018
111
Dear Sir/Madam,

I want to connect a 3V LED with a CR2032 (also 3V battery), why some experienced still recommends to add 50~100ohms resistor? What is the rule of thumb?

Best regards,

Kelvin.
 

sagor

Joined Mar 10, 2019
903
Many 3V LEDs are not exactly 3V, they may range a bit like 3.0-3.2V.
The idea of a resistor is to limit current. Even if you match a LED exactly to a voltage, it may draw more than the LED current rating. The resistor is just a safety feature to save the battery (and LED) from a short circuit current overload.
As a final thought, a CR2032 will not run a LED for very long, it has limited capacity. They are ok for very low current (microamps) applications, but not when drawing 10mA or so through a LED.
Those cells have about a maximum 500mAh capacity (most are in the 250mA range), so drawing 10mA would last less than 50 hours (down to about 2V) - not enough for your LED. Also, those cells are not designed for that kind of load, the recommended maximum continuous load rating is about 0.2mA.
 

Alec_t

Joined Sep 17, 2013
14,280
A '3V LED' is something of a misnomer. Firstly, the forward voltage Vf of a LED varies with the current through the LED, although for a white LED Vf is approximately 3V over a fair range of current. Secondly, due to manufacturing tolerances, no two LEDs have exactly the same characteristics, although they may be very similar. If you were very unlucky and had a LED with a nominal Vf slightly below 3V and a fresh CR2032 with a voltage slightly above 3V you could, unless you included a current-limiting resistor, have a LED current in excess of the safe limit for the LED. Usually though, the internal resistance of the CR2032 would limit the current sufficiently.
 

Audioguru again

Joined Oct 21, 2019
6,671
Nobody makes a "3V" LED. Some will be 2.8V, some will be 3V and some will be 3.2V or even a wider voltage range.
A new 3V battery connected to a 2.8V LED will destroy the LED and battery very quickly.
Also, the datasheet for the LED shows how the required voltage changes with temperature. I think it drops as it warms up which increses its current which makes it hotter which increases its current which …. it is called thermal runaway.
So of course you need a current-limiting resistor.
If the LED is actually 3.0V and the battery is 3.0V then you will see it dimming immediately as the tiny battery runs down.
 

dl324

Joined Mar 30, 2015
16,839
I want to connect a 3V LED with a CR2032 (also 3V battery), why some experienced still recommends to add 50~100ohms resistor? What is the rule of thumb?
For this combination, the current limiting resistor is typically omitted because the internal resistance and low capacity of the battery won't damage most LEDs.

This is the curve for Lx3340 LEDs:
1575738373845.png
If you extrapolate to 3V, the current will be under 100mA and these LEDs can tolerate 500mA briefly:
1575738456251.png
1575738497918.png
1575738522820.png

With the high series resistance of coin cell batteries, I wouldn't be too concerned about them killing these LEDs. But note that the maximum continuous current for these LEDs is higher than some.

Check the datasheet for your LEDs and do some experiments.

Discharge info for Energizer CR2032:
1575738853325.png
Even pulsed, the battery voltage will quickly drop below 3V.
 

Audioguru again

Joined Oct 21, 2019
6,671
Why was the graph of a 2V red LED shown when a 3V white, blue or modern green LED will be used?
Graphs are for a "typical" one that you cannot buy. You might get a 2.7V one or a 3.3V one, or worse.
 

dl324

Joined Mar 30, 2015
16,839
Why was the graph of a 2V red LED shown when a 3V white, blue or modern green LED will be used?
The OP didn't specify LED color.
Graphs are for a "typical" one that you cannot buy. You might get a 2.7V one or a 3.3V one, or worse.
You bring this up consistently, but, statistically, most devices in a bin are going to be typical. The outliers will be a small portion of the population.
 

Audioguru again

Joined Oct 21, 2019
6,671
I have a bag full of red Lumileds Luxeon name-brand LEDs. Their forward voltages are from minimum to maximum. It is difficult to match them so I measured the forward voltage of all of them then put them in "similar voltage" groups. Some LED manufacturers "bin" their LEDs into similar voltage groups.

I have a cheap Chinese flashlight with 24 white LEDs all in parallel. Sine the brightness of each LED is the same then somebody must have sorted through thousands of LEDs to group them to be the same. I have never seen Chinese LEDs that are already "binned".
 

BobaMosfet

Joined Jul 1, 2009
2,110
@Kelvin Lee

You want a rule of thumb: In electronics design, you _always_ control the quantity of electrons flowing via one method or material, or a combination of the two. For LEDs, you normally don't want to run them at maximum current-- somewhere a little lower is usually more than adequate.
 

MrChips

Joined Oct 2, 2009
30,706
Two months ago I had promised to write a blog showing why you need a resistor. I am halfway completing the blog hence that will have to wait.

The simple rule of thumb (as already supplied by other members above) is that no LED has a forward voltage rating of 3.0V. If the LED happens to be 2.8V and the coin battery is 3.2V you are in danger of blowing the LED.

If both LED and battery are spec'd at 3V there is no resistor value that will protect the LED and still give reasonable optical output.
 

oz93666

Joined Sep 7, 2010
739
Many cheap solar domestic 'garden lights' do not use resistor ... The rechargeable battery used is LiFepo4 ... this has an ideal , almost flat voltage of 3.2v , perfect for driving white leds direct ...

experimenting with these myself I find a 3W led driven directly from Lifepo4 consumes 400mA (recomended max is 750mA) ...as it warms up the current changes slightly .... under driving is ideal , less of a heatsink needed , more efficient , and lasts longer .... these batteries are tolerant of being taken low , which is also why they're used , although the led will naturally stop drawing current at around 2.5V so the battery will never go flat.
 

Audioguru again

Joined Oct 21, 2019
6,671
All the solar garden lights I have seen use an AAA Ni-MH cell producing 1.4V when fully charged and the 2V solar panel charges them with an average of about 35mA. They use a voltage stepup circuit using an oscillator, inductor and rectifier that provides about 15mA with high frequency pulses that limit the current.
 
For this combination, the current limiting resistor is typically omitted because the internal resistance and low capacity of the battery won't damage most LEDs.
Indeed. I've been playing around with some LED tealight (fake) candles recently. I was surprised to find they're nothing but a cr2032 and an LED, no resistor. Even shorted out, a cheap no-name new cr2032 only seems to be able to put out 80ma for no more than a second before it drops below 30 ma. The second most surprising thing about them is how long the "candle" can run and still look decorative in a glass holder. Days. And this with the steady on LEDs I swapped in because I didn't like the flickering ones.

No one has asked the op how long he needs his project to run, or what sort of light output he needs? It's not as if LEDs suddenly turn off below their rated current. They just get dimmer.
 
Many cheap solar domestic 'garden lights' do not use resistor ... The rechargeable battery used is LiFepo4 ... this has an ideal , almost flat voltage of 3.2v , perfect for driving white leds direct ...

experimenting with these myself I find a 3W led driven directly from Lifepo4 consumes 400mA (recomended max is 750mA) ...as it warms up the current changes slightly .... under driving is ideal , less of a heatsink needed , more efficient , and lasts longer .... these batteries are tolerant of being taken low , which is also why they're used , although the led will naturally stop drawing current at around 2.5V so the battery will never go flat.
How did you measure this without disturbing it?
A normal DVM on the current range (with probes) will probably have about half to one ohm of resistance. At 400 mA that will drop few tenths of a volt. When you disconnect the meter and reconnect the circuit the current will be higher than you measured.
 

Tonyr1084

Joined Sep 24, 2015
7,852
As a final thought, a CR2032 will not run a LED for very long, it has limited capacity. They are ok for very low current (microamps) applications, but not when drawing 10mA or so through a LED.
Last year [edit] in 2016 [end edit] I had several used CR2032's I wanted to completely deplete them before tossing them into the recycle bin. Decided to toss an LED with a Vf of about 3 Vf. The LED did not burn out. The batteries do not push that much current and I found that the LED's burned steadily for a couple weeks. I'm not sure if I kept photographs of the experiment, but the LED's didn't blow and the "USED" batteries lasted a long time.

I'm not talking engineering numbers because I just don't know them. All I know is from practical experimentation, used coin cells that could no longer open the garage door or ring the door bell still lit an LED without a resistor for weeks.

Give me time to review my older posts. When I find it I'll link it in.
 
Last edited:

Tonyr1084

Joined Sep 24, 2015
7,852
Found it. https://forum.allaboutcircuits.com/threads/coin-batteries-that-wont-die.129977/

This weird experiment went on from 12/9/2016 to 12;20/2016. Again, in January I ran another test. Initially I used 3mm superbright Blue LED's with an average Vf of 2.82. Minimum Vf measured was 2.80 and max was 2.83 with a standard deviation between all LED's tested was 0.01 Vf at one standard deviation. So the three blue LED's were all pretty close. The test I did in January used a 5mm superbright Red LED with an average 1.95 Vf. Since I used only one for this later test the Vf range is not important. But if you're curious, it was a range from 1.92 to 1.95 with a StdDev of 0.01 Vf.

My battery capacity tester has arrived and is presently employed on the task I bought it for. When it has a slot in its diary, I will be repeating the experiment and I will report the results.
Hey Al, did you ever get around to repeating my experiment?
 
Top