Transistor as voltage detector - Getting non-linear results

[Freddybaby]

transistor voltage detector

I'm using an LM395 (High reliability transistor) as a simple threshold detector. The source is half wave rectified AC (from a current loop) and through a resistive divider into the gate via a 5K resistor to prevent oscillation (per datasheet).

The saturation point seems very non-linear as I vary the pot. Coule the diode be the culprit ?

 

[millwood]

single transistor voltage detector

because the output is essentially the transistor's Ic/Vbe curve, as is.

what you can do is to introduce a little bit of negative feedback to linearize that Ic/Vbe curve: put a small resistor between the emitter and the ground.

how to size it? well it depends on your design objectives.

 

[Freddybaby]

threshold detector voltage

I got lazy and fixed this problem with a log function pot. Cheesy I know but it worked.. Thanks for the reply millwood