LEDs in Parallel != sum of expected current draw?

Thread Starter

mordhau5

Joined Sep 19, 2012
6
Hey guys, long time lurker, first-time poster!

I am attempting to create a loong string of non-HP, T-1 3mm LEDs as they would make great spotlights for an Amiibo (little plastic figures) shelf I'm building. I ordered 25 of these (3mm 3.3v 20mA: Datasheet here) from Digi-Key just to play around with their brightness, angle, and to prototype the string.

My plan was to run them 4-to-a-series, on a parallel string of probably 28 or so sets. I chose this configuration because
  1. I already have a large 12v PSU sending power to a set of RGB LED strips as a backlight to the display, and it should have more than enough capacity left-over for these little guys.
  2. running each LED at 3v provides AMPLE luminosity for the display and I may even have to include a pot to dim the whole circuit, but that's separate from this issue.
But the TLDR is: My concern is that when I pop 25 of them into my breadboard on the parallel power rails hooked up to my Dr. Meter benchtop PSU running at a fixed 3.0V, neither my standalone multimeter or the built-in Ammeter in the PSU read a draw of 25 * 1 Led at 3v (< 20mA). In fact, just putting two in parallel don't even come out to anywhere 2x the current. For example, I currently have 10 plugged in and I'm getting a draw of 114 mA but try any one of them in solo, and it reads 17-24mA. Given I'm a novice, but my understanding of components in parallel is that current draw is summed...right? I can for sure say that the brightness of the LEDs doesn't not seem to drop when more are added, even though their current draw seems to only increase by fractions of the individual led current. I even tried spacing them far apart just so their brightness' don't appear averaged to my eye.

I'm also to understand that sometimes low-quality benchtop PSU's can be unreliable at getting accurate draw readings (and Dr. Meter is certainly no hero...) but my multimeter seemed to confirm the phenomenon. I put this question to the measurement board because I have a hunch that I may be using my multimeter incorrectly to measure amperage and that my PSU's reading is wonky (it does seem to fluctuate when I turn output on/off). But honestly, I'm stuck! I don't want to proceed with ordering 100 more of these and wiring them up if they're actually drawing way more current than I'm currently seeing.
 

dl324

Joined Mar 30, 2015
16,846
Few of the members who have an electronics background would advise you to connect a lot of LEDs in parallel without a resistor in each string to deal with the fact that the LEDs will won't have matched forward voltages. In a worst case scenario, the LED with the lowest forward voltage would hog current and die. That could set up a cascading failure where the one with the lowest forward voltage hogs current and dies; until all of them are dead.

The human eye response to light is logarithmic and you'd need a change in brightness of about 2X to be able to discern the difference.
 

ericgibbs

Joined Jan 29, 2010
18,766
hi 5,
Welcome to AAC.
Note on the LED datasheet the Vfwd of the LED can vary from 3.3V to 4V.
So one LED could be 3.3V and another 4V and so on, they are not matched to operate exactly at 3.3V.

Putting a number of LED's in parallel means that some LED's will draw more current and other less.
Do not connect them in parallel.
E
 

Thread Starter

mordhau5

Joined Sep 19, 2012
6
Few of the members who have an electronics background would advise you to connect a lot of LEDs in parallel without a resistor in each string to deal with the fact that the LEDs will won't have matched forward voltages. In a worst case scenario, the LED with the lowest forward voltage would hog current and die. That could set up a cascading failure where the one with the lowest forward voltage hogs current and dies; until all of them are dead.

The human eye response to light is logarithmic and you'd need a change in brightness of about 2X to be able to discern the difference.
OH ok I see, so I've been thinking about this the wrong way. I need to treat the strings like each is their own LED with their own specs and pick a resistor that guarantees it will run at the same voltage as every other string in the parallel circuit. I suppose I would choose a fixed current on the PSU to allow them to draw then and see what voltage they require?
 

Thread Starter

mordhau5

Joined Sep 19, 2012
6
hi 5,
Welcome to AAC.
Note on the LED datasheet the Vfwd of the LED can vary from 3.3V to 4V.
So one LED could be 3.3V and another 4V and so on, they are not matched to operate exactly at 3.3V.

Putting a number of LED's in parallel means that some LED's will draw more current and other less.
Do not connect them in parallel.
E
I took those values to mean that any one of them COULD run at 4V but that would be outside of it's typical range and anything higher would run the risk of damaging the LED, but 3.3 was what they were optimized/QC'd to run at. I see now that it's the variance in manufacturing. Unfortunately running them all in series is simply not an option :( I'll need to find a way to get them in parallel with resistors.
 

bertus

Joined Apr 5, 2008
22,270
Hello,

It is not wise to parallel leds directly.
The leds will not have exactly the same forward voltages.
The led with the lowest forward voltage will draw most of the current and will fail as first.
Then the other leds will get more current and again the led with the lowest forward voltage will draw most of the current and then that led will fail.
etc. until all leds are dead.

Bertus
 

Thread Starter

mordhau5

Joined Sep 19, 2012
6
I'm already seeing that when I set the PSU to only allow a maximum draw of 15mA, the Voltage drop across each one varies somewhat, even the same LED in different slots on the breadboard. I think I may have what I need to proceed: take a fixed current draw and see what the required voltage drop is across each and use Ohm's law to pick a resistor for each series-set of LEDs that keeps them all at the same voltage drop (within some tolerance that allows for a long-enough lifetime of the string). I figure if I'm already keeping the current well below the rated amperage, then the small difference of increased current on the series with the lowest drop can still handle the difference for longer.
 

dl324

Joined Mar 30, 2015
16,846
I think I may have what I need to proceed: take a fixed current draw and see what the required voltage drop is across each and use Ohm's law to pick a resistor for each series-set of LEDs that keeps them all at the same voltage drop (within some tolerance that allows for a long-enough lifetime of the string).
It's simpler than that.

Assume a typical forward voltage of 3.3V for the LEDs. Three in series would drop 9.9V. The current limit resistor would be:
\( \small R = \frac{V}{I} = \frac{12V-9.9V}{20mA} = 105\Omega \)
Use 100 (or 110).

There will be small differences in brightness, but it shouldn't be noticeable.
 

Thread Starter

mordhau5

Joined Sep 19, 2012
6
It's simpler than that.

Assume a typical forward voltage of 3.3V for the LEDs. Three in series would drop 9.9V. The current limit resistor would be:
\( \small R = \frac{V}{I} = \frac{12V-9.9V}{20mA} = 105\Omega \)
Use 100 (or 110).

There will be small differences in brightness, but it shouldn't be noticeable.
That makes sense, but if I'm putting 4 in series, the required voltage drop to run at typical would be 13.2V meaning the limiting factor would be the 12V psu I'm assuming? Are you saying that you can't/shouldn't rely on your power source to provide no more than the rated Voltage? My original assumption was that because I'm increasing the drop beyond what the Power Source should be capable of, that that will limit my current on its own.
 

dl324

Joined Mar 30, 2015
16,846
Are you saying that you can't/shouldn't rely on your power source to provide no more than the rated Voltage?
A good quality power supply would provide the rated voltage from no load to full load. That's what I was assuming; 12V means 12V.

If you're using a low quality supply that gives a higher nominal voltage when lightly loaded, the voltage will drop when you get to the "rated" load.
 

Wiebenor

Joined Nov 22, 2018
7
Hmm... Digi-Key... i know that place. got a friend who works there in homeland security compliance, or whatever it's called. hes the guy that makes sure any parts sold, are not going to banned countries.

That's a good place to buy stuff I hear, although I haven't actually bought anything from them, as of yet. In your specific use though, would a set of white, or colored, strip lights, work just as well, i wonder?
 

dl324

Joined Mar 30, 2015
16,846
Digi-Key... [snip] That's a good place to buy stuff I hear,
DigiKey, Jameco, Newark, Mouser are all reputable. The only issue I have with DigiKey is high prices; though I've also heard that they have reasonable shipping prices. I haven't bought from them for decades; preferring Jameco and Newark.
 

Thread Starter

mordhau5

Joined Sep 19, 2012
6
Hmm... Digi-Key... i know that place. got a friend who works there in homeland security compliance, or whatever it's called. hes the guy that makes sure any parts sold, are not going to banned countries.

That's a good place to buy stuff I hear, although I haven't actually bought anything from them, as of yet. In your specific use though, would a set of white, or colored, strip lights, work just as well, i wonder?
I've been with them for a long time. I only order parts for rare project I work on, but whenever I need a component (especially since Radio Shack went out) I get excited to make an order from either them or Mouser. Picking the perfect parts is always made easy enough that even a novice can eventually find what they need, and they always get my orders perfect. Makes me feel a little like a manufacturer ordering for a big run haha.

I'm glad you asked because I've spent so much time trying to find one so I don't have to resort to this, as much as I love hobby electronics. I'm working with very tight tolerances because boards are only about .5" thick pine so tracks, datos, recesses (like for lights) have to be no more than 1/8" deep, so my spotlight LED recess is very small. I wanted to use those "fairy lights" you see all over amazon and boutique cafes, but they're essentially all made by the same manufacturer in China and they're just using insanely low power leds and too wide a viewing angle on the lens.
 
Top