Alright so mcd, or millicandela, is pretty much the intensity of the light emitted from an LED.

I'm making a UV meter with an LM3916...I know it has some comparators and some other circuitry inside, but would it be smart enough to adjust the output so the intensities are the same? The reason I ask is because I'm going to have green, yellow, red, (and maybe a blue) LED. I want them all to be the same brightness...and not have it like...green green green YELLOW! YELLOW! red red.

Assuming no...

Maybe this is subjective. I want LEDs that are bright enough to be clearly seen a dark room...but I don't want to light up the room with them, if you know what I mean. There's going to be a total of 20 LEDs. I don't want to go blind if they all get lit.

100mcd? 1,000mcd? 10,000mcd?

Also, how big is the difference between say an 80mcd LED and a 60mcd LED? So if I can't match the LEDs perfectly, how far away can they be from each other without being noticeably different intensities?

Thank you very much.

 

Forget about millicandelas. They aren't a valid unit since they include a dispersion angle and don't take into account the varied sensitivity of the human eye to color. You can convert the millicandela and viewing angle data into lumens but really you're going to just have to adjust by eye, especially since your application is not about transmitting light to illuminate a surface but for direct viewing.

http://en.wikipedia.org/wiki/Candela
http://en.wikipedia.org/wiki/Lumen_(unit)

Another note about millicandelas. Often a manufacturer will offer 2 or 3 different viewing angles using the exact same die mounted in different die cup (reflector) geometries. The effect is to direct the available energy into a narrower or wider beam (viewing angle) while having an inverse effect on intensity in millicandelas. The illumination measured in lumens will be approximately equal for each.

Did you mean UV meter or Vu meter? For direct viewing, a good diffuser layer goes a long way and careful choice of color helps. For example, the typical green LED is 570nm wavelength that looks quite yellow. A better green LED emits at 520nm and is more pleasing.

I'd recommend experimenting with various LEDs using an LM317L regulator set up as an adjustable current regulator using a potentiometer. Once you've determined the highest current needed, you can program that into your 3916 and then offset the other colors with a series resistor for each. For example, let's say you've chosen a red, yellow and green LED and when you drive each at 15mA, the yellow is dimmest. Then you would adjust the current from the 3916 to suit the yellow LED and use an experimentally determined series resistance with the red LED and another with the green LED to make them appear the same intensity as the yellow.

 

Convert to lumens?

 

Hi BigJason,

KJ6EAD makes some excellent points. The mcd rating is important, but it has to taken with respect to the viewing angle. That said, if the goal is to clearly see the LEDs without lighting up a dark room, diffused LEDs rated 100mcd or less will be ideal. Someone can correct me if I'm wrong, but diffused LEDs have a wide viewing angle, so as long at the mcd is under about 100, you shouldn't have to worry about viewing angle being a factor.

If you needed to see the LEDs indoors with windows or lights on, then 50 to 400 mcd would be ideal, again diffused.

Also, diffused LEDs are often the color of the light they emit. So most diffused red LEDs are red for instance.

What type of LED are you thinking about? Round (3mm or 5mm diameter) or rectangular? Rectangular ones are often already diffused and used for meters like the one you're building. I don't know if they make a blue rectangular one though. I know they do in 5mm and probably a 3mm. I can direct you to a handful of suppliers once I know the type you'd like.

Hope this helps.

 

Oh, also, you can sand the LED surface to diffuse the light better and/or change the series resistor to get more or less light.

What voltage are you working with?

How many of each LED color are you planning to use, i.e. two reds, two, yellows, three greens, etc.?

 

Oh sorry I meant Vu meter. I was just watching a video on UV (lol).

Alright. So by using an LM317 in adjustable current, it's essentially working as series resistor that can be adjusted? Once the LED is the brightest it can, then I measure the current using a DMM.

Then I program the LM3916 via the resistor. The current should be for the LED that needs the most current to get to full brightness. The others which may be brighter than the one just selected can be dropped via a series resistor, like a single LED.

Fair enough. I think I've got that right, (do I?). And then how should I figure out the resistance of the series resistors?


I plan on using 5 green, 3 yellow, and 2 red on each. There will be two. Voltage can be picked. This will be in an amp that runs off of 35VDC. I can regulate that down to whatever. I was thinking either 12 or 15V to the IC, if that's what you meant.

Thank you.

 

So by using an LM317 in adjustable current, it's essentially working as series resistor that can be adjusted? Once the LED is the brightest it can, then I measure the current using a DMM.

Then I program the LM3916 via the resistor. The current should be for the LED that needs the most current to get to full brightness. The others which may be brighter than the one just selected can be dropped via a series resistor, like a single LED.

Fair enough. I think I've got that right; do I?

Yes.

And then how should I figure out the resistance of the series resistors?

Assuming you'll be breadboarding this first, leave a space for the resistor and jumper it with wire until you've finished everything else; put a low value potentiometer in series and dial-a-brightness, then measure the pot value and swap in the nearest fixed resistor. You may find that all LEDs will look good at the same current if they're from the same manufacturer and model.

 

Yeah I was planning on using the same series/make for the LEDs just for some uniformity.

By low value pot, as in like 1K? Or as in like 50ohm? This might sound silly, but I don't really want to have to make another order after I've determined what value the resistors need to be. I think I might just use trimmers. I have a ton of them. I DON'T have a ton of fixed value resistors.

Thanks. I appreciate it.

 

By low value pot, as in like 1K? Or as in like 50ohm?

50Ω or 100Ω should be good.

The formula for converting millicandelas to lumens is included in the Candela Wikipedia article but using it only provides an approximate value since the formula is for a cone of consistent luminous flux but LEDs have a lambertian dispersion pattern. You can calculate using the 2θ1/2 (half power angle) value from the LED data sheet or you can use one of the online calculators such as this to get the same result:
http://led.linear1.org/lumen.wiz.

 

50Ω or 100Ω should be good.

The formula for converting millicandelas to lumens is included in the Candela Wikipedia article but using it only provides an approximate value since the formula is for a cone of consistent luminous flux but LEDs have a lambertian dispersion pattern. You can calculate using the 2θ1/2 (half power angle) value from the LED data sheet or you can use one of the online calculators such as this to get the same result:
http://led.linear1.org/lumen.wiz.

Thanks! I guess I'll go about ordering everything.

To find out what series resistor I should use, I'll use ohms law, I think.

 

To find out what series resistor I should use, I'll use ohms law, I think.

Yes, subtract the LED forward voltage from the power supply voltage and that gives you the voltage for the resistor, with that and the recommended LED current it's easy to work out.

 

Yes, subtract the LED forward voltage from the power supply voltage and that gives you the voltage for the resistor, with that and the recommended LED current it's easy to work out.

Thanks. I know how to calculate the series resistor in typical situations...but since I have a programmable current limit on the LM3916.

I'm going to use an LM317 in current limit mode and then find out what current I want. So would I use the current I set...not the LEDs forward current? Or..?

Thanks.

 

The LED's rated current is just a maximum above which damage can occur. The LED will start to emit light at just a few mA.

 

The LED's rated current is just a maximum above which damage can occur. The LED will start to emit light at just a few mA.

Thanks.

I'm tired...but I just realized I'll also need some sort of preamp because 0.3-0.4V isn't anywhere near 10V (which is what the National schematic is for).

So I'll get a 5532 or something and set a gain of 20 or 25 maybe...right? RMS and peak to peak confuse me.

 

Alright so mcd, or millicandela, is pretty much the intensity of the light emitted from an LED.

I'm making a UV meter with an LM3916...I know it has some comparators and some other circuitry inside, but would it be smart enough to adjust the output so the intensities are the same? The reason I ask is because I'm going to have green, yellow, red, (and maybe a blue) LED. I want them all to be the same brightness...and not have it like...green green green YELLOW! YELLOW! red red.

Assuming no...

Maybe this is subjective. I want LEDs that are bright enough to be clearly seen a dark room...but I don't want to light up the room with them, if you know what I mean. There's going to be a total of 20 LEDs. I don't want to go blind if they all get lit.

100mcd? 1,000mcd? 10,000mcd?

Also, how big is the difference between say an 80mcd LED and a 60mcd LED? So if I can't match the LEDs perfectly, how far away can they be from each other without being noticeably different intensities?

Thank you very much.

The difference is of about 20mcd LED the difference won't be that big it might be negligible

 

The difference is of about 20mcd LED the difference won't be that big it might be negligible

It's always a good idea to check the dates on a thread before replying to it so you don't look foolish; this thread is absolutely ancient, and the person you're giving advice to hasn't even been on AAC in 9 years. You are, in effect, talking to yourself in public.

 

Oh, also, you can sand the LED surface to diffuse the light better and/or change the series resistor to get more or less light.

What voltage are you working with?

How many of each LED color are you planning to use, i.e. two reds, two, yellows, three greens, etc.?

Are LED lights available in neon colours?

 

Yes.

 

Lumens. The splendor of light is normally estimated in lumens, or the aggregate sum of light produced. ...

Candelas. You may see some LED lights estimated in mcd, or millicandelas. ...

Lux. Another estimation of light is lux, or the light estimated in 1 square meter of yield zone.

 

Alright so mcd, or millicandela, is pretty much the intensity of the light emitted from an LED.

I'm making a UV meter with an LM3916...I know it has some comparators and some other circuitry inside, but would it be smart enough to adjust the output so the intensities are the same? The reason I ask is because I'm going to have green, yellow, red, (and maybe a blue) LED. I want them all to be the same brightness...and not have it like...green green green YELLOW! YELLOW! red red.

Assuming no...

Maybe this is subjective. I want LEDs that are bright enough to be clearly seen a dark room...but I don't want to light up the room with them, if you know what I mean. There's going to be a total of 20 LEDs. I don't want to go blind if they all get lit.

100mcd? 1,000mcd? 10,000mcd?

Also, how big is the difference between say an 80mcd LED and a 60mcd LED? So if I can't match the LEDs perfectly, how far away can they be from each other without being noticeably different intensities?

Thank you very much.

They're evaluated in millicandelas, or thousandths of a candela. Millicandela is ordinarily truncated as mcd. Common power of a T1-3/4 (5mm) red LED is around 50 to 75 mcd. On the other hand, purported superbright LEDs may have a force of 500 mcd to more than 5 candelas.