Minimizing Vbe variation by controlling Ic?

[tinska.h]

Hello,

I'm building an accurate current limiter with limited amount of potential difference in voltage supply. I'm using a N-channel MOSFET as driver and then having a NPN BJT to do feedback when current exceeds desired amount. Circuitry looks a bit like this: **broken link removed**

The problem is that because feedback circuitry's turn on is based on NPN transistor's Vbe. And I get Vbe variation from device to device, which then affects my current limit accuracy. And the situation gets worse when I need to use small resistors.

In transistor's datasheet it suggests that Vbe difference in different temp. environment is smaller when Ic is higher. Could I possibly assume that when I have higher Ic, my Vbe variation from batch to batch would also be smaller?
Any other suggestions how to minimize Vbe's effect? Differential amplifier could be an option, but don't want to go with matched pairs.

Thanks,

 

[crutschow]

How "limited" is the voltage?

For more accuracy and stability you could add an op amp and accurate reference voltage in the loop to take the Vbe variations out of the equation.