Hi everyone,
I am doing project involving photodynamic therapy on cancer cell lines using led.For that purpose i have to create arrays of led (i.e 96 LEDs in 8*12 fashion) so that equal current flows in all branches of led and there will be no current fluctuation.... i have seen many circuits with LM317 but i needed one with transistor....so any sort of answers will be highly appreciated...I'll be waiting for reply....!!!....Thanks....!!!

 

What do you think an LM317 is besides several transistors in one, convenient package?

If you want to do this with one transistor at a time, we will need to know the supply voltage you have available, the voltage of the LEDs, and the current rating of the LEDs.

 

Need to know:

LED part number?
What current do you want to drive through the LEDs?
Available power supply voltage?
Ambient temperature and variation?
Cost Budget? (Build drivers or buy)?

 

If there 12 branches of 8 LED, you'll need 12 pieces of LM317; if there are 8 branches of 12 LED, you'll need 8.

If you want to use discrete components, you'd still need an op-amp, pass transistor and zener diode for each branch. Here is what is inside the LM317...



[edit] see post#12 for the correct block diagram.

 

...If you want to use discrete components, you'd still need an op-amp, pass transistor and zener diode for each branch...

Not quite: Follow this link. Look at post #7. Circuit can be scaled to lower or higher currents. Also look at post #15 in that thread.

 

If you put all 96 LEDs in series it matters not what type of driver you use: each one will have exactly the same current as all the others with mathmatical certainty.

Just be careful with the >250 volt drive required.

 

Not quite: Follow this link. Look at post #7. Circuit can be scaled to lower or higher currents. Also look at post #15 in that thread.

Yes, you're right. I have read this thread some time ago but cannot recall where I read it.

I knew IRFZ44N is a high current low voltage mosfet, so it is a very good choice for this purpose. If the LED is running at 330mA then may be he can put more LEDs in one string and increase the voltage to 24V.

Allen

PS: your circuit is an elegant design.

 

[QUOTE="absf, post: 890941, member: 108875"Here is what is inside the LM317...[/QUOTE]

The 317 does not have a ground pin. This looks more like a 78xx diagram.

ak

 

Need to know:

LED part number?
What current do you want to drive through the LEDs?
Available power supply voltage?
Ambient temperature and variation?
Cost Budget? (Build drivers or buy)?

In addition to this ^ please give dimensional requirements on the size of the sample/work area (intended illuminated area),etc..
Also max lumens required.

Depending on the size/lumens you may be able to make your life much easier with a single "higher" powered LED or premade LED cluster/array..
Example.. Maybe you just need 1 x 50W LED

Also state if you need the ability to "dim" these LEDs (change their intensity)
Also state the desired color temperature or wavelength required..

Personally if I was doing something "cancer/medical" related I'd want full intensity control,exact spectrum/wavelength,etc... aka.. fancy stuff
And I sure has heck wouldn't want to cobble 96 gumdrop LEDs on a perf board with multiple hacked LM317 circuits/wire wrapping/sloppy solder jobs,etc...

With the right information it could be as simple as an off the shelf dimmable driver and potentiometer for intensity control.

 

What do you think an LM317 is besides several transistors in one, convenient package?

If you want to do this with one transistor at a time, we will need to know the supply voltage you have available, the voltage of the LEDs, and the current rating of the LEDs.

What do you think an LM317 is besides several transistors in one, convenient package?

If you want to do this with one transistor at a time, we will need to know the supply voltage you have available, the voltage of the LEDs, and the current rating of the LEDs.

Thanx for suggestion...do you mean that i can get LM317 with transistor in a single convenient IC...As i already said i wil have 12 branches of LEDs and 8 LEDs in each branch...then the voltage in each branch will be around 8*2.15(assume)=17.2 V and current i want is around 20 mA through each branch...and i am using red and green leds...

 

In addition to this ^ please give dimensional requirements on the size of the sample/work area (intended illuminated area),etc..
Also max lumens required.

Depending on the size/lumens you may be able to make your life much easier with a single "higher" powered LED or premade LED cluster/array..
Example.. Maybe you just need 1 x 50W LED

Also state if you need the ability to "dim" these LEDs (change their intensity)
Also state the desired color temperature or wavelength required..

Personally if I was doing something "cancer/medical" related I'd want full intensity control,exact spectrum/wavelength,etc... aka.. fancy stuff
And I sure has heck wouldn't want to cobble 96 gumdrop LEDs on a perf board with multiple hacked LM317 circuits/wire wrapping/sloppy solder jobs,etc...

With the right information it could be as simple as an off the shelf dimmable driver and potentiometer for intensity control.

Thanx...!!!...I need 96 leds in 12 branches with 8 leds in each branch..so the voltage in each branch would be 8*2.15(assume)=17.2 V and current to be around 20 mA in each branch....I am considering simple focused leds till now....And yes definitely i want intensity to be controlled and intensity varies with current so current needs to be regulated..So there must be equal current flow 20mA in all branches of LED and there should be no current fluctuation....And i am using red and green leds....But there is not specific intensity i require only optimum glow of LEDs anyway 50J/cm^2 will be good....and dimension of exposure is (12-13cm)*(8-9cm) which is very small space....anyway should i consider high power leds....i have never worked on them..do they provide high intensity even for small current.....?????

 

Here is what is inside the LM317...]

The 317 does not have a ground pin. This looks more like a 78xx diagram.

ak

You're absolutely right.... Here's the correct one. Looks more complicated.


Allen

 

The LM317 can be configured as a current source, but all of the output current goes through the sense resistor, varying the output current is messy, the headroom voltage requirement is relatively large, and overall it is just messy. Try this:

12 parallel branches x 20 mA = 240 mA, well within the capability of a single LM317.
One current limiting resistor per string. Rcl, 1/4 W - 1/2 W.
24 VDC regulated output -17.2 V string voltage - 6.8 V. for approx. 20 mA, Rcl = 331 ohms (1% tolerance) and Iled = 20.54 mA
Vary the LM317 output voltage in the usual way to vary the string current and brightness, with low current resistors and pot.
Power dissipated in the LM317 at 24 Vout = 0.246 x (Vin - 24) W.

A smaller current limiting or ballast resistor can increase efficiency, but it does less to even out variations in string brightness caused by small variations in the Vf of each individual LED.

ak