Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations GregLocock on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Constant current LED driver

Status
Not open for further replies.

RTaylor

Bioengineer
Mar 28, 2003
1
Hello, I am a neuroscientist with a nodding acquaintance with electronics (and no in house electronics support). I need to use an LED to stimulate light sensitive neurons in the retina. The current needs to be stable so that the light source can be accurately calibrated. I plan to drive the LED in two modes: 1) as a continuously variable voltage from a DAC (0-5V), 2) from a oscillating TTL output (the idea is to linearly adjust intensity by adjusting the logic high duty cycle).
I imagine the following specifications would be suitable:
1. Current stability <1%.
2. Max output current ~50mA.
3. Rise-time time-constant < 2usec, i.e. bandwidth ~500kHz.
Does this sound feasible?
As I will be building this myself, any general or specific suggestions for circuits, and their possible pitfalls would be greatly appreciated.
Or does someone know of a commercially available device?
Thanks in advance,
RTaylor.
 
Replies continue below

Recommended for you

Dear R:
I assume you want white but I don't know if color changes with current/age/temp.

Controlling the voltage is not a good way. Voltage drop
is non-linear with current, changes with temperature.

+/-1% current is easy, but light efficiency may change
with temp.

Suggest: Use contstant current, pulsewidth modulated
and use a photodiode with sample/hold to control
the amount of light.

Can you tell me more details?
<nbucska@pcperipherals.com>
 
LEDs are diodes. Diodes are square Law devices. That means that brightness does not change linearly with current.
The voltage applied to the LED is not relavant as nbucska has stated. You must limit the current to the allowable values of the LED used. The forward voltage drop of the LED is function of color and composite material. Check your datasheet.
 
You could use an adjustable voltage regulator such as the LM317 to make an adjustable constant current source. Here is a link to a schematic:
However this aproach requires a variable resistor to change the current level.

You may be able to switch the current source on/off with a transistor. However I don't know if the LM317 could be turn off and on as fast as you want (500kHz). I suspect that the light intensity would be more proportional using the duty cycle aproach.

If 24 mA is sufficient then you could have your DAC feed a AD694 chip which will convert it to a constant current. Data sheet here:
 
Dear Mr. Taylor,

I can design the circuit for you and email it to you if you send me your email address. You can email me at cappels @iname.com. No charge, of course.

Dick
 
Beware of temperature/efficiency variation.
The duty cycle is Ton/Tper Both can be varied, so the resolution may be reduced.
<nbucska@pcperipherals.com>
 
Note that Steve Owens has a point when you consider the accuracy you are asking for. You may need to resort to fiber optics and a fiber optic splitter to implement the feedback. The feedback control is not trivial because of response time considerations. Another approach could be to use a calibration scheme with a table lookup to drive the LED. In this case, you would drive the LED with something like a 12 bit D/A and create a lookup table using a known light sensor. Once calibrated, your driving circuit would use the table lookup to output the current needed for the LED for each desired output. Note also, as stated before, temperature should be held constant between calibration and application. 500KHz is not a trivial problem either and may eliminate the use of a feedback system for reasonable cost.
Good luck.

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor