Current Limter Design, constant voltage

[KerimF]

I like to point out that there are two kinds of current limiting protection.

(1)
If the load tries to sink more current than the limit, the output voltage drops while keeping the maximum current almost constant.
In this case, if the regulator is linear, the excess voltage (the drop) will be developed (added) on the IC regulator terminals (Vin-Vout) hence it increases its heat dissipation. But perhaps this is an advantage if this overheating can shut off the output current completely as it is likely the case for LM317 (To be honest I hadn't the chance to test this in real), though the time delay may not be acceptable at the load side.

(2)
When the load tries to sink more current than the limit, the output current is cut off completely by a sort of latch up circuit. Obviously this is more complex than the previous one.


In any case, you did well in getting some ICs of LM317 (it is one of my favorite regulators). It has a short circuit protection too (I have just noticed that I didn't intentionally test this function yet... you will I guess :smile: ). So you may need just a simple current limiter.

I would like to suggest that you find out the maximum acceptable current (as a normal load in your application) then the current limit would be adjusted to be twice of this maximum (to have a good margin just in case). This will make the solution more reliable and practical.

Kerim

 

[ALERTLINKS]

Current Limter Design, constant voltage

As i am an old timer i just recalled an application which was used to serve exactly the same purpose, it even does not use any active component. A miature reed switch is wound with few turns of wire is put in series to the load acting as current operated relay. I remember four or five turns of 1mm wire, when checked for continuity with multimeter caused reed switch to close which in turn operates a latching relay cutting off the supply.