LEDs in Series Question

[LyleCaldwell]

OK, I have a mixer with a 30 VDC power supply.  I tried running 6 LEDs (ultra bright blue water clear LEDs, Mouser part number 604-WP7104PBC/H) in series after a 10K resistor.  According to every online LED resistor calculator, this should have been more than enough - 320 Ohms would be the "safe" rating.

Based on another project I did where I liked the light value from this model of LED when supplied with 9 VDC using individual resistors per LED, I *knew* that a 15K resistor would be ideal, but that a 10K resistor (what I had on hand) should be just a bit brighter.

But instead I find that of the 6 LEDs, the first and third (measuring from where the 30 VDC + is attached to the first LED) LEDs are very bright, while the second and fifth LEDs are noticeably dimmer, and the fourth and sixth LEDs are even more dim.

Now, I only chose to use series LEDs because I "could" with the 30 VDC supply, and to save some wiring time.  I don't mind going back and tapping just 15V of the supply and doing individual LEDs.

But I'm curious why I'm getting the uneven LED brightnesses.  All 6 LEDs are brand new from the same supplier, and I've never had any variations using these LEDs before in projects.

Thoughts?

BTW, I learned from this to test LEDs *before* using hot glue on them.  DOH!  Lots of fun awaits me.

[stm]

AHAAA!!!!

This is one of the most fascinating physic phenomena I have witnessed lately.  I've been experimenting with the newer GREEN and BLUE LEDs, those with forward drop voltages above 3V.

In one case I put a 6 strings of 12 series blue LEDs (yep, 72 blue LEDs). Each of this six strings was fed from a 42V supply with a variable resistor for brightness control.  Indeed at low currents some LEDs just turn off before others!  However, I tested all LEDs before installing them and were fully bright.  I changed the positions of the LEDs and the ones that got dim first kept doing so.  I took them out and tested individually and worked at full brightness with currents above 5 mA approx.

Then, I made an experiment with a string of 8 series GREEN LEDs and at low currents some of them started blinking with an irregular rythm (in average at a 1 sec period).  For currents higher than 5 mA no problems were observed.

The veredict:  I read on a datasheet some brief comment about operating these LEDs near nominal current, but no additional explanation.  My wild guess is that these devices exhibit a sort of tunnel effect at low currents, and as such when a certain threshold is crossed (downwards) they almost turn off. Remember light emitting phenomena has to do with electrons changing between certain well defined states or "orbits".

Hope someone can provide deeper and more supported insight to this phenomena.

In your particular case, I deduct you are applying a current around 1 mA to your LEDs (30V - 6x3.2V)/10k = 1mA.  (BTW, I am assuming 3.2V forward drop and not the typical 3.5V drop because you are using low currents, and the latter value is pecified for nominal current, i.e. 20mA).

One experiment I haven't tried is powering the string form a CONSTANT CURRENT SOURCE instead of a simple resitor and supply.

Best regards.

[amz-fx]

I think STM has hit the problem on the head...    the very low current is causing this....

So,  change the LED to 4.7k and see how they look.  If that isn't okay, go with 2.2k and you should definitely notice an improvement.

regards, Jack

[LyleCaldwell]

I thought about that.  Believe me, I've been thinking about anyting that will let me just swap out the one resistor instead of digging through the hot glue globs.   

The trouble is, unless these superbright blue leds are fed between .5 and 1 ma (values I've used when running one resistor per LED), they turn into little blue flashlights.  At 20ma they turn into little blue spotlights.

I need to see if RS sells little 10K trimpots - I don't want to pay shipping getting a good one from Mouser.  I have some 100K trimpots but they are too abrupt in the lowest range of the track.  Otherwise I'll just run a few diffferent resistors I have on hand to find the right value.  I have a 3K and a 1.5K on hand - I'll see how the lights do at 4.5K (though I suspect they will be too bright - I'll see if I can find the threshold where the lights start to be uneven).

[LyleCaldwell]

Well, I finally got all of them to light up evenly.  Using a 330 ohm resistor.  So of course I have six little spotlights.  Any greater value makes the LEDs have different output levels, with those closest to the + DC supply being brighter.

So I'm going to write off using series LEDs with these Super Bright blues, and put in six individual resistors.  Darn it.

[Mike Burgundy]

You could supply them with a chopped supply (short high-frequency bursts of voltage - vareying the on time varies apparent brightness) but that has a somewhat higher parts-count (ahem), and might cause trouble with your audio circuits. But it's possible. 

[Paul Perry (Frostwave)]

I think the problem must be, that unless the leds are perfectly matched, they develop a different voltage across them for a particular current, so some glow before others. Also, you have heating effects, so it could make a difference which one turns on first. Maybe this effect is worse for blue leds.
That's the good thing about electronics, always something new to bite you on the ass...