LPD8806 Pixel Anomaly / Power Bleeding (Packaging Issue?)

[bamboosamurai33]

Has anything come up with this issue?
I bought my third roll and I am having the same issue here. Just with this one.

I did notice that the third "resistor" "I think" from the beginning of the strip looks a little dinged up almost. I noticed it before I even hooked it up as I was removing the rubber wrap. Im wondering if the first driver chip is in a funk. I am going to add a segment just before the first one, so it wont be the first in line. See if that works.

[adafruit]

we havent been able to determine whats up yet, its quite mysterious!

[Jelmer]

Any new insights on this problem?

I have a string of 370 leds divided in 5 segments and each segment is connected with a cable of about 1m until about 10cm. I see the first green led on every segment light up according to the the clock speed of the data I clock into the string.
I also have some connections of about 1 to 2 cm and i can't detect any bleeding there. So it must have something to do with the length of the connection cable connected to the input of the string.

If I can find some time I'll try to compare the signal before and after the first driver on each segment and see if I can find a difference.

[Jelmer]

Forgot to mention that I have 3 different set-ups of this kind and I observe the problem with all 3.

[pounce]

What code are you using?

[Jelmer]

My own implementation of the protocol on MBED.

[pounce]

Do you have the ability to test the strip using arduino and the out of the box test implementation? As you can imagine its sort of hard to tell you if you have a hardware issue if you are using your own code etc.

[cjbaar]

FYI, I thought I would jump back in here.

pounce... if you refer back to the beginning of this thread, you will see in my second post where I show that this is definitely a hardware issue, and is independent of any code/libraries used. I was able to duplicate the issue by doing nothing but pulsing the clock pin (code sample is shown there). The more the clock pin is pulsed (as in a PWM duty cycle), the brighter the green LED gets.

Also note that (at the recommendation of pburgess), I was able to mitigate the issue by placing a small resistor in the clock line. This doesn't solve the issue, but it makes it less noticeable. In my case, a 330ohm worked, but it seems to vary depending on the length of the line I used.

[pounce]

I was responding to jelmer. This person has 4 posts and using their own code. I don't assume their issue is your issue when there are unknowns.

[ScottishDave]

this is definitely a hardware issue, and is independent of any code/libraries used. I was able to duplicate the issue by doing nothing but pulsing the clock pin (code sample is shown there). The more the clock pin is pulsed (as in a PWM duty cycle), the brighter the green LED gets.

Just to +1 this, I've found this thread after having the same 1st led flicker issue. I have tried driving the strips (I have 1m and 2m strips) with an UNO r3, a Meag2560 R3 and a mini - all with the same result. I happen using just about everything I run, and I always use the adafruit LDP8806 SPI library.

I use the strips for photography and it does show up sometimes on my images, so a proper fix would be great, although I appreciate its not the end of the world.

[gzumwalt]

I too am experiencing the dimmly lit green led in the first position. Is there a sollution?

Thanks,

greg

[adafruit]

we haven't yet discovered why this occurs or how to avoid it. we suggest simply covering the first pixel with opaque tape if its a problem until we can determine where its coming from.

[JohnSan]

I can add a bit of info for this.

I have split a 5m strip into its 10 sections.
you can see the strips are manufactured in a 16 led section.
The first led in each strip is on a very dim green.

Prior to the strip being separated, I'm pretty sure I would have noticed all those when testing' it, if they were doing it then. So I'm pretty sure they weren't. As for just the very first led, I probably wouldn't have noticed. The connection lead length at that time was what comes with a 5m strip. I took the socket off the trailing end of the strip and used it to connect to the Arduino 'terminal' shield to plug into the strips, leading end plug. So maybe 0.3m length?

But now.....
The first strips', first led, is a 'brighter' dim green than the others. (The one connected to an Arduino Uno).

My strips are pluggable but still form one strip of 160leds. (So DI/DO CI/CO are still daisy-chained).
Whichever strip is plugged into the connection directly from the Arduino is the one with the 'bright' dim green led. So probably not manufacturing hardware issue.

It seems to be associated with the length of the CI/DI & CO/DO wires and occurs when show() is called.

If the strip is not being written to, the green glow is not present. (So they are either all doing it or none are).

Since the DO/DI & CO/CI connections between my strips are now nearly one metre and the Arduino connection to the first strip is approx 2m, I think it's cross coupling type of interference. But whether its related to 0V or just the signals themselves, I don't know.

I've done all my testing with this LPD8806 strip, just using the USB port power supply from my laptop plugged into the Uno, which also feeds the strip.
So I've monitored current at times and its quite a bit lower than you would expect and has not caused any apparent problems. (I'm quite consious of how may 'white' leds are on at a time).
However, I have tried two completely separate power supplies to feed the led strip independently- just 0V connected to Arduino. Neither of which had any effect on this.

I've tried supply capacitor decoupling etc, to no avail.
But this issue is no big deal for my application at the moment.

Just information...

[garymuhonen]

Has anyone found a solution for the green flicker on the first led of the lpd8806 led strips?

The green flicker appears when the line length from the micro to the leds is over two feet long.
When looking at the signals on a scope, there is a lot of ringing on the data and clock lines.

I made a matrix out of these (24x20) and the first led in each row has the green glow, since there is about 30 inches of wire between each strip.

I've tried some simple things like adding a capacitive and/or resistive load on the inputs to the led strips, but didn't find an easy solution.

[mohawkpiper]

just saw this thread...
first, this led strip was my first time ever using arduino and leds, so i didnt think much of some issues when they came about, maybe cuz i didnt know what to think, or what to compare it to.
Anyhow, this happened to me also.
First, Im using a teensy 3.0, running 3 meters, but in testing, it was a build up from 4 LEDs to the total 96 LEDs. I was using the example code for LEDbeltKit which slowly got modified as i went.

So,
I had cut every single one of my LEDs into the 2 LED sections, and started wiring up my project. (I needed them all spaced apart much further than they come in the strip.) I noticed this green LED light very slightly upon first testing. thought maybe a bad LED, switched the cut section to another, same thing... it didnt matter what LED it was, it had happened regardless. It was always the first one in the chain. I tried quite a few of the cut sections. I just figured some power was making it to the LED itself when it shouldnt have, but in my project the LEDs are mounted inside a plastic bike frame, and this "glitch" in no way could be noticed through the plastic so i shrugged it off and continued with my project.

Now, I only remember this glitch in the early stages of my project, halfway through i split the data and clock wires to run two separate strips at the same time.
i just opened the frame now and took a look, i dont see this glitch anymore. I dont see it on the first LED of either of the strips. I dont know when this went away and i dont know why.

Unfortunately everything is mounted and i cant try to reproduce this, but i feel this rules out the "maybe the first LED in a reel is bad" kinda thing mentioned earlier.
Also, my wires from the teensy to the first LED were never longer than 6 inches, but currently are about 2 or 3 inches.

Dont know if this helps at all.

G