Hi,

    I want to thank Michael and Dwayne a lot !! I am still too much a
digital guy :-( It took me a whole page to really get and see how the
polarization of the transistor base with the diode could work ! It is a very
nice option for projects like mine that need to limit current at about 4
amps. I will be able to use smaller resistors and get dissipation way down.

> >     I need slower vbe than silicon to use in a short circuit protection
> >system. The 0.6 volts of silicon transistor makes me use a large wire
wound
> >resistor that could have it's value divided by 3 and would also get the
> >dissipation much lower.
>
> You can bias the base of the transistor so that it is not quite turned
> on.  There are a couple of ways of doing this - how temperature stable
does
> this whole thing need to be?

    It is ok, in my application, to have up to 15% error in the trigger
point. I think I will not have much problems if I use the diode drop as
Michael pointed out. VBE will change a little, the diode will have
variations and the resistor divider should also contribute to variations,
both static caused by components variations and dynamic dominated by
temperature changes but the 15% limit seems to me as easy to do even with 5%
resistors. Am I not seeing something there ?

> The easiest way should just involve adding a single resistor.  You
probably
> have the base connected to your current sense resistor via a 1K or so
> protection resistor.  Just add another resistor to the base that pulls the
> base towards conduction by 0.4V or so.
>
> The words sound confusing but it is very simple.  It would be easier if
you
> could describe your circuit.  Or even better - draw your circuit in ASCII
> and post it.

    I think I clearly understood it, after the one page brainstorming :-) .
The diode solution should have less temperature variations and they should
move together with the variations in the transistor VBE, so it should be a
nice solution.

    I am awfully bad with ASCII circuits :-( Roman's program do not run on
my machine. The circuit is basically what Michael has drawn without the
diode and resistor to ground. I think I am allright using the polarization
diode and making the base stay at about -0.4v . I will protoboard it in the
next days and post the results. If it works as I think it should it is a
much cheaper solution than "smart fet's" and those expensive right side
current monitors.

Best regards,
Alexandre Guimaraes

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.