APA102 2020 Brightness specs
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

APA102 2020 Brightness specs

by tjwatson on Tue Jun 06, 2017 11:12 am

I have a question about where Adafruit got the brightness and power specs for these LEDs (https://www.adafruit.com/product/3341 for reference).

The webpage lists the brightness specs as:
Red Brightness @ ~20mA: 600-800 mcd
Blue Brightness @ ~20mA: 800-1000 mcd
Green Brightness @ ~20mA: 1500-1200 mcd

But the datasheet lists:
Red Brightness: 300-330 mcd
Blue Brightness: 420-460 mcd
Green Brightness: 160-180 mcd

Did Adafruit perform their own brightness checks independent of the datasheet? It wouldn't surprise me to find out the datasheet was wrong, just curious where your numbers came from.


Posts: 1
Joined: Thu Mar 23, 2017 10:51 pm

Re: APA102 2020 Brightness specs

by AhiP on Fri Aug 23, 2019 1:25 am

What do you mean? The web page and datasheet report the same numbers, it may have been updated since your question though. Or, if indeed sources claim different mcd values, here's why I think... well why.

TLDR: There really shouldn't be ambiguity in mcd measurements. It's best got from some intrinsic properties of the LED that the manufacturer would know:
LED's max power consumption (W), times its luminous efficiency at that point (lumens/W), yields max Lumens. To get mcd, scale that by how wide the view angle of the LED is. The manufacturer would and should have those values and probably are the ones who made the datasheet (which I hope everyone is referring to on their sites).

If we were to try and get mcd without access to such data and measure it ourselves, it gets complicated:

mcd stands for "millicandela" as in the datasheet

Candela (Italian for candle probably Lol) as far as I know measures the power a light source emits, normalized by how much of your field of view the source covers (really it's how much of the light you're blocking, but they're proportional and that's useful in this example). The closer you go to the LED, the more radiation is reaching you (as you physically cover more of it), but the light source also makes up more of your view, resulting in an mcd that's unchanged. Once close enough (you cover all of the light rays) if you were to focus it (into a beam), the same power is reaching you but the light source covers less of your field of view, and mcd goes up.

Problem is measuring power emitted by a light source directly is difficult. Much of the radiation won't be picked up (at least accurately) by the light meter sensors leading to a large margin of error when generalizing from those readings (as the intensity varies across view angles). It's a lot easier to measure Lux (how much a surface is lit up, in this case the sensor catching the radiation), and convert it to mcd. Problem with this method is it requires some subtle but specific requirements such as measuring from somewhere on the axis of the light cone, that is, you facing it directly. Unless you've a high-end light meter that calibrates and compensates for this, it'll be a very inaccurate reading. This leaves us with the only viable choice on a budget, that is to measure brightness per area directly in Nit (which still takes a fairly expensive meter actually). Since Nit is how much Candela per area, we cannot use 1 LED as that's a point source and has no area (it only appears to have an area because of the protective layer diffusing the light). So we'd have to use an LED panel instead (an area that's consistently lit) and divide by the number of LEDs in the panel to get the average mcd per LED.
See a problem? We could have different densities of panels, different wiring (powering sectioned groups will have less voltage drops, and more consistent brightness across the LEDs than a string would) etc. etc.

Ultimately what I'm saying is it's best to contact the manufacturer and do some math, than to buy an expensive device, and then do some math to calibrate it.

Posts: 1
Joined: Thu Aug 22, 2019 9:56 pm

Please be positive and constructive with your questions and comments.