PNP current limiter accuracy

[treez]

Hello

sorry if this is too many on the same topic-please delete me if you have seen enough of this

I am doing a LED lamp powered by SEPIC SMPS.

It is 10W and the LEDs will be end up being switched in one by one by FETs, so that the lamp comes on “chaser” style.

Here is the schematic

http://i25.tinypic.com/qoziat.jpg

Anyway, since any hold-up on the output caps may cause a LED overcurrent after switch-on, we have a hi-side PNP current limiter in the circuit.

Hi side PNP current limiter close up

http://i31.tinypic.com/2jd25uw.jpg

Anyway, am I right in saying that this PNP current limiter would not be able to allow us to set an accurate current limit, and that the particular current limit would depend on the particular characteristics of the particular diodes and transistor that are used in it ?

So we would not be able to guarantee say a 330mA current limit since the diode Vf’s and the PNP transfer characteristics would vary from part to part.

Can you say whether or not the above shown current limiter would allow us to accurately set the current limit?

 

[benta]

Realistically you can land within 5...10%, but not with that circuit.
Why so complex?
Remove the two 1n4148s, they are unnecessary. Replace the 1n750 with a 3.3 V Zener. Calculate the emitter resistor for a voltage drop of 4.0 V.
BTW, it would be nice to be able to read the voltages in your circuit.

Benta.

 

[benta]

Ooops! 4.0 V should read 2.5 V.
Editing is not possible here?