A single op-amp antilog amp?

[<<marjan>>]

> >I have seen the basic topology for an antilog amp - diode in place of the
> >input resistor in an otherwise standard op-amp amplifier - but I don't
> >remember how to do the math to calculate the resistor values. Obviously the
> >input voltage to the antilog amp must be limited to a range of 0 to about
> >0.7 volts - but how do I calculate the current through the diode (and hence
> >the feedback resistor)?
> 
> how indeed....
> you canput a pot in series with the diode & connect across a battery
> (perhaps a fixed resistor as well to protect the diode0 and then measure
> I and V across the diode as you move the pot settings... then plot it all
> and you will be the wiser.... this device will be very temperature
> sensitive, BTW.
> 
> Would like to hear people's opinions on this. There is  table in "The Art of
> Electronics" for zillions of different diodes, if you are looking for a high
> impedance (likely) try a diode made by connecting the base to the collector
> of a
> transistor, possibly.
 
I've seen in theory something like this.It said that the better version
of that
diode thing is to put transistor instead diode,npn C to (-) of opamp,E
to output
and base to gnd.
Also better temp comp was made with that diode/opamp then another diode
to out of
opamp ( K to opamp out) ,some precise Iref between Anode and gnd
(current to diode)
and that to input of noninverting opamp which has PTC in place of
resistor in 
feedback to gnd leg.It's Av must be high ( than 1 ) .
Vout= - const x (Av noninv) x ln ( Vinput / Rinput x Iref).
Well,just in theory...



marjan