Re: [Help-gnucap] Model for led diode

[al davis]

On Monday 01 September 2008, Jesus Genicio wrote:
> I make the change in "is" and ok, the diode led run in 2.46
> volts with 5 volts in Vin but de vce with vin=0 is only the
> 2.8 volst.....

> gnucap> tran 0 5 0.2
> #Time       v(Vin)     v(D1)      v(Rc)      vce(Q1)
>  0.         3.2074f    2.1327     163.09n    2.8673
> ..............
>  2.6        5.         2.5594     2.3628     0.077836
> .................

>
> Vin=0 => Led off but v(d1)=2.13 Volts and led light. Sorry,
> the "is" ok but there are some parameter more that configure
> to make that diode run correctly?.
>
> One reply is that " you need more study of theory of
> circuits" :)), please be pacient. Thanks.
>
> the only way that i found for this, is:
>
> vin=0, V(d1)=2.13 but ic=163nA and this is off.?

Check the current  ...  You can probe it, but my calculator says 
I(Rc) = 652 uA.  The LED won't light up on that.  If you think 
of it like a resistor, you could substitute a 3.2 giga-ohm 
resistor for a similar result.

Below the knee, a diode is essentially open, so there can be 
some voltage there with nearly no current.  With a little bit 
of leakage in the transistor, some extra due to the 
simulator's "GMIN", .. There could easily be some volts across 
the LED.

If you try it on a real circuit, you would get a similar 
situation, until you hook up the meter.  Then all of the 
voltages change.  A typical meter is equivalent to a 10 meg 
resistor.

If all you care about is the LED, it is probably ok as is, but 
not a good design.  You should add a "pull-up" resistor, from 
the collector of the transistor to the power supply, or in 
parallel with the LED.  I think 100k is a good value.