[sdiy] Log amplifier circuits

[Oren Leavitt]

On 1/18/2012 1:39 PM, Harry Bissell wrote:
> Thanks to all who replied and sent links or attachments. I found several circuits that do exactly
> what I expected.  The overall results are virtually identical (I expect that with some more value tweaks, they would
> be identical.
>
> I also think I remember a circuit using the LM13700 that was supposed to be a log amp (but is really exponential). I'll
> have to mine the data sheet and see if I might try something with that (it could be temperature compensated, at least good
> enough for me... and I'm using other LM13700 packages
>
> H^) harry
>
>
> ----- Original Message -----
> From: Harry Bissell<harrybissell at wowway.com>
> To: synth diy<synth-diy at dropmix.xs4all.nl>
> Sent: Mon, 16 Jan 2012 11:37:19 -0500 (EST)
> Subject: [sdiy] Log amplifier circuits
>
> Hello all. Playing with LTspice I used a ln (natural logarithm) math function and it worked
> well for me. I tried to make a log amp using a diode or transdiode configuration, but the dynamic
> range is much more limited than the math function was.
>
> I tried to play around with the input resistor value (more of less current into the log amp) but that
> didn't have much effect.  It seems like the values are limited to voltage ranges right around the
> forward drop of the diode(s) or transistors.
>
> anyone have any ideas or circuit references for me to look at ?
>
> tia
>
> H^) harry


Hi Harry, It can be a royal PITA to get more than 3 or 4 decades of good 
log/anti-log out of these things.
Sort of like keeping marbles from rolling off of a wobbly table.

Sergio Franco's "Design With Operational Amplifiers And Analog 
Integrated Circuits" has some good discussion on the matter, as well.

- Oren