LM317 current limit circuit blew up

[betwixt]

Looks GROSSLY over-engineered to my eyes.
I think I could do this in maybe 15 components in total. I'm thinking an MCU to do all the timing things and an op-amp with current sensing to maintain constant current.

Brian.

 

[Xenobius]

Looks GROSSLY over-engineered to my eyes.
I think I could do this in maybe 15 components in total. I'm thinking an MCU to do all the timing things and an op-amp with current sensing to maintain constant current.

Brian.

Could you direct me to a schematic to see please?

 

[Easy peasy]

my quick anaylsis shows the LM317 will die every time you pull the adj pin to gnd for ~ 30V output ( 480mA will be drawn )

 

[betwixt]

Could you direct me to a schematic to see please?

I do not have one, it was my guess at what would be needed if I had to solve the same problem.

The way I would approach it is:
1. join as many LEDs as possible in one chain, it simplifies the constant current circuit because the current will be the same through every LED if they are in series.
2. connect a transistor (MOSFET or BJT) at the bottom end of the LED chain. One transistor even if more than one chains are connected.
3. add a resistor (~1 Ohm) between the source/emitter and ground.
4. use an op-amp to control the conduction of the transistor based on the drop across the resistor, this gives you the current control.
5. for timing, use a simple MCU to produce the reference for the op-amp and switch it on/off as needed.

If more LEDs are needed than one chain allows, just parallel the chains with a small current balancing resistor in each.
This method uses far fewer components and dissipates much less heat.

Brian.

 

[danadakk]

You could always eliminate the OpAmp and just use MCU A/D to do the current
measurement off the Isense R. Simple control loop. And many MCU have SAR
which translates to fast decision, in uS range, of over current which generally
most components can handle fairly short over current transients.


Regards, Dana.