Hello all you enlightened people, Back in February,

Archive FastLED Archive
1

Hello all you enlightened people,

Back in February, I was just about ready to start on my dream of a 16X16X16 cube. Well… that suddenly came to a grinding halt and has been on hold ever since… :frowning:

I have now kind of revived that project but decided to go for a significantly smaller 8X8X8 size for now. My dream cube will simply have to wait…

Now, I have all my components on hand, Teensy3.1, OCTOWS2811 adapter, 150W 5V PSU, rolls of 20AWG solid tinned copper wire and a bag with 600 8mm RGB LEDs with embedded WS2811 chips and miscellaneous ‘doo-dads’ to complete the thing.

Now, I want to individually test those 600 RGB LEDS to pick 512 (and some…) with very uniform brightness levels for all 3 colors.
Real duds (If I find any…) should be easy to screen out but I am wondering about potential brightness variations (Chinese manufacturing you see… ). I want to do this before I start soldering them into that cube as I know that any repairs will be a serious pain !!

So I would like to automate these tests and need to select some photo resistor or photo diode circuit that is sensitive enough at all 3 colors to detect the smallest possible variations in brightness. Any suggestions here ??

Thanks!

2

What about one of these products? https://www.adafruit.com/categories/61

3

Adafruit has a couple sensors that might do the job for you. Here is one http://www.adafruit.com/product/1334

4

Hi Jon, Herb, thank you very much !!

The TCS34725 seems a good choice but also noticed the TSL2591 that have a higher dynamic range. Would I need that much for an actual light source !? (RGB LED as opposed to reflected light )

Will dig in the specs a little bit more and will select soon as I am very eager to start this build.

Maybe I’m being overly cautious with that level of testing but I am willing to wait until I have 100% confidence in all the RGB LEDs I end up using.

5

Yes, your test rig will have to be able to maintain a constant distance from the sensor for every LED you test. How precision are we talking here? I like the idea however.

6

@Jon_Burroughs yes I agree, I already have an idea of a jig to achieve this constant test LED to sensor distance.

To be honest I do not know how much precision I need…

I intend to drive the test LED through all values (0-255) for each of the R,G,G colors. Collect the measured sensor output in a table and compare that with the next LED.

Eventually, I will get some standard deviation that could be used in a GO NOGO test for the rest of them!?

Even if I do not find any flaky LEDs, it will at least eliminate the doubts I have about that cheap Chinese manufacturing !

7

@J.P_Saini , I guess it all depends on how much variation there is but you may very well be right, it will not matter et the end !!!

I am also now thinking that due to the LED’s PWM control of brightness, it may be tricky to correctly sample the brightness with these sensors !?

I need to make sure that the sensor is actually not so fast as to possibly sample only part of a pwm cycle.

8

You might want to look also at

The spectrum response for R and G is equal, B is 50%. In addition the sensor has an ALS as well as direction detection what makes it interesting for automation.
You should go for White-balance and brightness.
However, binning is quite complicated:
http://www.cree.com/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/XLamp%20Application%20Notes/LED_color_mixing.pdf

You also might contact @Phil_Xu , maybe he can point you in the right direction where to get already pre-binned LEDs .

9

Thank you @Juergen_Bruegl for that technical reference on LED color mixing !

That device you suggest however does not quite compare with the ones I have seen from Adafruit.

It is too late for me to order any pre-selected RGB LEDs. I already have them on hand and only want to select 512 (plus a few spares) out of my order of 600 that are as similar in brightness as possible.

10

As suggested by Jon and Herb, I Just ordered the Adafruit’s RGB Color Sensor - TCS34725. Hope to receive it soon and start putting together a valid automated test sketch to verify each of the 600 8mm RGB LEDs with integrated WS2811 I ordered to build my 8X8X8 cube (512 RGB LEDs).

Once I sort out the test jig details and Adafruit’s TCS34725 library example sketch parameters, I hope to find out the actual variation in color and brightness from that batch of RGB LEDs.

I will come back with test results a.s.a.p. and will make that test sketch available. Maybe someone else (unfortunately not you strip users :wink: could benefit from that effort.