Probably making a stupid mistake, but need someone to point it out to me

centizen:
In my experience, without a current limiting resistor, LED's simply burn out in a couple of seconds. I was under the impression that these ones didn't burn out when I applied 3.3v because they had a built in resistor to the LED. I can't remember who I heard that from, but looking back it's not too plausible.

Well live and learn. Leds almost always require external current control. So unless you find a datasheet that explicitly shows some form of internal current control then it's on your design to provide it. You don't power leds with voltage, you control the amount of current that is allowed to flow through them.

Lefty

So you think that the reason this isn't working is because I'm missing current flow control? i'll try building another one tomorrow with resistors on it if that is the case.

centizen:
So you think that the reason this isn't working is because I'm missing current flow control? i'll try building another one tomorrow with resistors on it if that is the case.

Without a proper schematic drawing I'm not prepared to say why it isn't working. However I will say that those led do require external current control.

Lefty

I have some LEDs like these that I got from the Digikey. They are a bit temperamental in terms of how they're biased. I think your problem is that the blue and red cathodes aren't entirely isolated from each other. The datasheet doesn't actually give pin information (annoying, and annoyingly common for RGB LEDs...) so it's possible you're not even using the right pin as the anode. Any way to verify you've wired them right?

I know I've got the wiring right on that board for each respective LED. But that makes a lot of sense actually. Are there any good RGB LED's that you know of that won't have this bias issue?

Since you don't have any current limiting, then the power supply and circuit wiring is doing it for you; otherwise the LEDs would have burnt right out. The forward voltage drop (Vf) of the red LEDs is less than that of the blue ones. Since you effectively have them in parallel the voltage is too low to turn the blue ones on once a red one is plugged in. Take a volt meter and see what your power supply voltage measures at the circuit board, I bet it doesn't measure what you expect when the LEDs are on.

centizen:
I know I've got the wiring right on that board for each respective LED. But that makes a lot of sense actually. Are there any good RGB LED's that you know of that won't have this bias issue?

No I don't actually. I haven't used many. I'm always surprised that they don't work like three single LEDs sharing the same plastic. For e.g., both RGB LEDs I have on hand refuse to light up multiple colors at the same time. You have to PWM between the R, G, and B legs sequentially if you want to do any color mixing.

However, none of my blue lights turn on, only the red ones.

Despite the assumptions, despite what the datasheet may or may not be appearing
to say, the only way to make any sense out of this whole thing is to take a single
Led in hand, and play with it until you get the pin connections straightened out.

A lot of DMMs have a diode test position, and it would take 30 seconds to figure the
whole thing out.

Then, I would take 5V and a 1K resistor and see how the Leds glowed.

Then, if I truly believed they had built-in Rs, I would take exactly one Led with the
intent that I might be sacrificing it to the gods of thermodynamics, and apply full
voltage to it, and see if it glowed for an instant and then forever flamed out.

ONLY THEN would I hook up the original ckt shown. Better than blowing up $18 worth
of Leds at square one.

Lastly, if I were really really desparate, I might go read where adafruit mentions
something about "three legs to ground through a resistor".

SirNickity:

centizen:
I know I've got the wiring right on that board for each respective LED. But that makes a lot of sense actually. Are there any good RGB LED's that you know of that won't have this bias issue?

No I don't actually. I haven't used many. I'm always surprised that they don't work like three single LEDs sharing the same plastic. For e.g., both RGB LEDs I have on hand refuse to light up multiple colors at the same time. You have to PWM between the R, G, and B legs sequentially if you want to do any color mixing.

That sounds like you are only using one resistor per LED.

Looking forward to the OP's schematic being posted on this one.

oric_dan:

However, none of my blue lights turn on, only the red ones.

the only way to make any sense out of this whole thing is to take a single
Led in hand, and play with it until you get the pin connections straightened out.

Then, if I truly believed they had built-in Rs, I would take exactly one Led with the
intent that I might be sacrificing it to the gods of thermodynamics, and apply full
voltage to it, and see if it glowed for an instant and then forever flamed out.

ONLY THEN would I hook up the original ckt shown. Better than blowing up $18 worth
of Leds at square one.

Yeah, I did all that. That's how I know that the LED works at full current on 3.3v, at least the 3.3 volts supplied by my Teensy3. That's also how I know I've for sure got the wiring right for this particular model of LED. I know that I am not going to blow these out at this voltage, and my question is why the blue lights aren't turning on when the red ones are on, not whether or not I'm about to blow up 20 bucks of hardware. I'm not.

However, thank you for the recommendation for the three grounds and resistor on Adafruit. I'll be making a better board today and implement protection and such so that when I'm running it at high voltages they don't burn out.

Grumpy_Mike:
Looking forward to the OP's schematic being posted on this one.

Oh boy... I'm not very good at schematics, but I'll try. Give me a bit.

Oh boy... I'm not very good at schematics, but I'll try.

Make them look like you see the pros do. Have the positive lines at the top, the ground at the bottom and signal flow if any from left to right.
Have a look at some of the schematics on my site.
http://www.thebox.myzen.co.uk/Hardware/MIDI_Footsteps.html
and others, linked at the top of the page.

That's also how I know I've for sure got the wiring right for this particular model of LED.
......... and my question is why the blue lights aren't turning on when the red ones are on,

The 2nd half of this comment indicates the 1st half is not entirely accurate. All in all,
electronics really is magic, but hooking up Leds is not.

Um, I must respectfully disagree. I know which pin is the common anode, and which ones are the ground for each respective color. What other information should I know about this? The lights work fine on separately, and when all are set to one specific color, and I can flash one between blue and red and green no problem in quick succession. Is there something I am missing here? Because I don't see how this is relevant when we established much earlier in the thread that my issue is probably due to a lack of current control or lack of forward voltage bias separation between the LED's that I have. Sorry if I'm coming across as rude, I'm just genuinely confused.

centizen:
Um, I must respectfully disagree. I know which pin is the common anode, and which ones are the ground for each respective color. What other information should I know about this? The lights work fine on separately, and when all are set to one specific color, and I can flash one between blue and red and green no problem in quick succession. Is there something I am missing here? Because I don't see how this is relevant when we established much earlier in the thread that my issue is probably due to a lack of current control or lack of forward voltage bias separation between the LED's that I have. Sorry if I'm coming across as rude, I'm just genuinely confused.

Did you ever measure the voltage at the LEDs? I really suspect that the device driving the LEDs (Arduino pin) was doing the current limiting for you and dropping the voltage down to the Vf of the red LEDs when one or more of them was on. As soon as they were all off, the output voltage would rise to the Vf of the blue LEDs. I don't know about the mega specifically (since I haven't looked into it yet), but many microcontrollers have approx 100 ohms of output impedance. This automatically limits the current to 50mA max.

I'm just genuinely confused.

And so are we. What you report is not possible so there is something wrong that you are not telling us. We have had some guesses and you say it is not that so we are waiting for your schematic to throw some light on the situation.

Yep, I'm just at work at the moment so once I get home I'll be making a schematic. Any particular recommendations for a software package to do that with? Otherwise I'll go with ExpressSCH.

My very first test with these pancake RGB LEDs was with a multimeter on diode/continuity mode. I use that to test LEDs a lot, since it's high enough voltage to light one up, but is current-limited enough that it's not particularly bright.

So, with one of these LEDs in hand, I tested all the pin combinations (my data sheet was a little more helpful, but not much) to find the one pin that seemed to be the constant required to illuminate all of the colors. Great. So, then, I tried shorting the R and G pins to make yellow. Nope. The R and B pins to make purple. Nope. The B and G pins to make cyan. Nope. If I set the tip of my probe on one color and rocked it over to touch another, the first would be lit up until I touched the second, which would stay lit until I rocked the other way, then the first would come back on.

I think Grumpy is on to something here. I bet what is happening is that the lower forward voltage diode is pulling the supply down enough that the higher forward LEDs are no longer lit. Isolating each internal diode with current limiting resistors would probably fix that . . . Huh, glad I stopped by this thread.

centizen:
Any particular recommendations for a software package to do that with?

It is called pencil and paper, photograph the results and post.

thank you for the recommendation for the three grounds and resistor on Adafruit.

Three things. First, you need 3 separate resistors, one in each Led cathode line, in
case you're not doing that.

Secondly, if you check the datasheet [which is one of the worst I've seen yet - they
show no schematic and 2 A's and 2 C's, rather than 1 A and 3 C's like adafruit says],
you'll see the d/s mentions needing 3.8V to turn on the green Led, so it won't light if
you have only 3.3V.

Thirdly, the d/s says 3.8V for green, but it's usually the blue that takes the most voltage
to turn on.