Control Multiple Leds Brightness

[slide]

I am in need of a way to control approx 10 LEDs brightness (all the same brightness) but I can't use PWM because these LEDs are used to illuminate a small area where a Linear Array will sense the reflection. The LED drivers I've seen online seem to be much more than needed. Can anyone help me figure out a low part count solution?

 

[kender]

How much current do you need to drive the LED with?
By the way, what kind of experiment is this hardware for?

 

[slide]

My leds have a peak forward current of 100mA so with 10 leds 1A max.

These will provide lighting for a Linear Sensor Array made by Taos. The sensor can detect at a 5MHz rate so thats why I need steady brightness.

 

[kender]

How about a current regulator like on p.17 of this datasheet: **broken link removed**

?

 

[slide]

I need some way to control the brightness from a uC

 

[solidstate]

Maybe you can feed back the sensor array's output to control brightness. If your signal is at AC (you mention 5 MHz), create a low pass filter with corner frequency well below that (few Hz maybe), that should output a near DC measure of the optical input to the sensor. Compare that to a reference and let the difference drive a current or voltage regulator.

Using a uC will require AD and DA steps in between..

Would this be an idea?

 

[kender]

I need some way to control the brightness from a uC

Recently I've posted a controlled current sink circuit. I'm using it for friving LEDs. Look in this thread . It's PWM-controlled, bu the LEDs see linear voltage - they don't pulse to PWM. That circuit can be DAC-controlled too (PWM with a low-pass filter in that circuit is just a poor man's DAC).

Keep us posted!

- Kender

 

[slide]

solidstate:

This is the description of the Linear Sensor Array (Taos TSL208R) that I'm using:

The TSL208R linear sensor array consists of a 512 × 1 array of photodiodes and associated charge amplifier circuitry. The pixels measure 120 μm (H) by 70 μm (W) with 125-μm center-to-center spacing and 55-μm spacing between pixels. Operation is simplified by internal control logic that requires only a serial-input (SI) signal and a clock.
The TSL208R is intended for use in a wide variety of applications including mark detection and code reading, optical character recognition (OCR) and contact imaging, edge detection and positioning as well as optical linear and rotary encoding.

The 5MHz I mentioned is how fast you can clock it. After each clock tick the next photodiode reading is placed on the output of the device. I am using this to detect a black line on a white background.

I know how big the line should be, so the μC will be able to calculate if its detecting too much black (not enough light) or too little black (too much light). This way it can adjust the leds output to get the correct width reading. I don't think an analog method will work for this.

kender:

Anyway to make this drive 10 or so LEDs at once? That seems like it would work for 1 but I need many.

 

[kender]

kender:

Anyway to make this drive 10 or so LEDs at once? That seems like it would work for 1 but I need many.

Connect the LEDs in series - all 10 of them.

 

[slide]

I only have 3.3V out from uC or approx 11.1V straight from battery.

 

[kender]

I only have 3.3V out from uC or approx 11.1V straight from battery.

Connect them straight to the battery. Without knowing anything about the device you're building or experiment you're preparing or illunination intensity and uniformity requirements, I'd say - use a smaller number of larger LEDs.

By the way, Light/current ratio for the LEDs will have some batch variation. Light/current ratio will vary appreciably with temperature.