maximize ir led strength

Forum 2005-2010 (read only) Hardware Interfacing
1

hey there,

for ir-tracking (wiimote) purposes i'm trying to pulse ir-leds to maximize their light-output. i'm not sure if i already reached the possible maximum yet. maybe you can have a look over my specs and code?

i have to run two leds: SFH 415 (HTTP 301 This page has been moved)

the datasheet says, that their forward voltage should be around 2.3V when pulsing them with t = 100 microseconds. this will give me I = 1A.

my wiring is like that and i use a transistor to drive the leds

my code looks like this:

int ledPin = 8;

void setup()
{
  pinMode(ledPin, OUTPUT); 
}

void loop()
{
  digitalWrite(ledPin, HIGH);
  delayMicroseconds(100);
  digitalWrite(ledPin, LOW);
  delayMicroseconds(100); 
}

am i done? :slight_smile:

2

for ir-tracking (wiimote) purposes i'm trying to pulse ir-leds to maximize their light-output.

I wouldn't do that, a Wiimote is basically a camera and pulsing the LEDs is going to be a problem because the camera will miss the light if it is not on when the camera scans that part of the image.

Also 10R is too small to regulate the drive current in an LED, there is too much room for error due to variations in forward voltage of the LED. Look at the data sheet and you will see a range. can the LED handle 1A peak current anyway?

3

thanks for your thoughts.

apparently i'm experiencing problems tracking pulsing leds when they move to fast. i'm thinking about alternately pulsing two leds. when pulsing with 50% duty cycle this should look like one strong led for the camera.

the led datasheet says that the forward voltage should be smaller than 2.8V, ideally 2.3V when pulsing with t=100us. the current then will be 1000mA.

for me it seems that i don't need any resistor at all?

thanks for your help

//edit

just found a standalone circuit using a 555 (via Pulse Circuits for Infrared LEDs and Visible Diode Lasers)

Many infrared GaAlAs LEDS can be driven at currents approaching 1000 mA provided that the duty cycle and pulse width is short enough to keep the average power dissipation low enough. Using a standard 555 timer IC and a 2N2222a NPN transistor, it is possible to create a drive pulse with width down to a few microseconds and amplitude to about 800 mA. The 555 timer chip has a rise/fall time of about 100 ns and current drive capability of 200 mA. The transistor output driver raises the output drive capability by about a factor of 4.
The diagram below shows an implementation of this approach. The 555 timer circuit uses 1N914 switching diodes to enable complete control of the on/off time for the pulse. Despiking and power supply decoupling capacitors of 0.1 and 5 uF were used. A 220 ohm series resistor in the 2N2222a base circuit is used to set the base drive level. A 10 ohm collector resistor provides suitable current limiting. The LED in the emitter circuit is shunted with a parallel 100 ohm shunt resistor to speed up the fall time of charge drain. The 555 and transistor were biased at the same adjustable level with a common power supply (22-121 Micronta Adjustable Dual Tracking DC power supply rated at 1 Amp and 15 V).

4

What Grumpy_Mike is explaing is the LEDs should not be pulsed, but should be on all of the time.

If you have a look at Johnny Chung Lee's Wiimote stuff:
http://johnnylee.net/projects/wii/

You'll see that he uses Vishay TSAL6400 LEDs running at 100mA continuous.

You could do approximately the same with the LEDs your using.

5

Many infrared GaAlAs LEDS can be driven at currents approaching 1000 mA provided that the duty cycle and pulse width is short enough to keep the average power dissipation low enough.

Yes but this extra light is of no use to you for tracking with a camera precisely because it is flashing. You get no advantage in this arrangement flashing a brighter light or having one or two lights on continuously. The average light output is still the same and that is the output that matters.

6

thank you for your comments!

i checked some of the sites you mentioned and ordered some leds from farnell: TSAL5300, TSAL6400, SFH4231, VSML3710

especially the VSML3710 look interesting, because the have a continously forward current of 1A. http://www.farnell.com/datasheets/86271.pdf

7

I am headed down the same path of pulsing IR leds to get a brighter output for the purpose of improving the range on a wiimote camera.

@hans12000, have you pursued this further?

@Grumpy_Mike, I'm still confused why pulsing LEDs would not be beneficial. If multiple LEDs are pulsed in a sequence to overdrive them, and those LEDs are clustered together, then from a distance wouldn't the wii camera see one continuous light, brighter than a normal led?

Hmm.. maybe this is the scenario?:
If you overdrive an LED to 10x brightness, does that mean it can only be on for 10% of the time, and off for 90%?
If that were true, then it would take a cluster of 10 LEDs to cycle through for a continuous light. If those 10 LEDs were instead replaced with 10 normally driven LEDs, would the brightness be equivalent?

8

If you overdrive an LED to 10x brightness, does that mean it can only be on for 10% of the time, and off for 90%?

Yes

If those 10 LEDs were instead replaced with 10 normally driven LEDs, would the brightness be equivalent?

Do the sums. If you only have it on for one tenth of the time AND you are feeding it with 10 times the current, then that is the equivalent of an LED on continuously with unitary current. So pulsing 10 LEDs in sequence at 10X the rated current would be equal to just 10 LEDs at normal current. But then you wouldn't have to have to have a high current switching device on each LED and something to time that they went on in sequence.

9

Thanks for the reply GM!

Another variable I'm not smart about here is how the wiimote camera works.
I assume it has a kind of IR CCD. I'd assume the wiimote reads the CCD values periodically. Lets say 100Hz. The "average light" argument suggests that the CCD is actually accumulating IR readings continuously, between each of the 100 read requests each second. And that it is not just getting IR reads on demand (100Hz).
I hope that makes sense? Is it a continuous averaging or a single read as needed? It may be a stupid question... but I don't know how CCDs work.

10

It is a built in sensor and processor:-

11

I still don't know how it works internally. So how do we know how often it samples IR light? That seems critical to knowing if pulsing would be advantageous.

12

So how do we know how often it samples IR light

You don't.

That seems critical to knowing if pulsing would be advantageous.

Yes exactly.

13

When you look at the diagrams in good datasheets we will also find that the brightness will not be proportional to the current!!

14

So what's the bottom line here? Is pulsing and/or overdriving an IR LED potentially advantageous to get better tracking with a wiimote camera?
By better tracking, I mean better distance or better detection in ambient IR conditions (outside).

I'm understanding that if I pulsed my LED when the camera was sampling it would see a brighter target.
And that pulse may require less average power consumption than a continuous light(s).
And it would require fewer LEDs. (though more electronics)

When you look at the diagrams in good datasheets we will also find that the brightness will not be proportional to the current!!

So you're saying that you may use less (or more) average power when pulsing to get the same light as when not pulsed.

15

He is saying that more current may not produce proportionally more light, it will drop off at the higher current values. So a current pulse of 10 times the standard current will produce less than 10 times the light.

So what's the bottom line here?

I think we are saying we don't know but it is not as easy as you think. You could always do some experiments and see. My guess (it is a guess) is that pulsing the LEDs will be more trouble than it is worth.